Praxisbeispiel: Data Science im Marketing

Wie Sie mit Data Science die Conversion-Rate in Ihrem Online-Shop erhöhen

Die Fragestellung: Ein Hersteller von Elektrogeräten lancierte einen neuen Online-Shop, um einen neuen Vertriebskanal zu schaffen, der unabhängig von stationären Einzelhändlern und Amazon ist. Obwohl der Online-Shop von Interessent:innen häufig besucht wurde, war die Conversion-Rate zu niedrig und der Umsatz somit zu gering.

Die zentrale Frage war nun: Wie kann die Conversion-Rate erhöht werden, um den Umsatz über den neuen Vertriebskanal zu erhöhen?

Was ist eine Conversion-Rate? Die Conversion-Rate ist eine Marketing-Kennzahl, die in diesem Beispiel das Verhältnis der Besucher:innen des Online-Shops zu den getätigten Käufen meint. Halten sich viele Besucher:innen im Online-Shop auf und sind die Warenkorb-Abschlüsse dennoch gering, so ist die Conversion-Rate niedrig. Das Ziel ist es, die Conversion-Rate zu steigern, also dafür zu sorgen, dass Besucher:innen, die sich im Online-Shop befinden und dort etwas in den Warenkorb legen, auch einen Kauf tätigen.

VorgehenUm zu verstehen, warum eine Bestellung abgeschlossen bzw. nicht abgeschlossen wurde, wurden verschiedene Daten aus dem Web-Analytics-System des Online-Shops untersucht. Dazu gehörten im Wesentlichen Daten zu Besucherhandlungen auf der Website, die automatisch getrackt, also aufgezeichnet werden, wie z. B. Button & Link-Klicks, Bildergalerie öffnen, Produktvideo ansehen, Produktbeschreibung ausklappen, Time on page usw.

Mit diesen Daten wurden drei Analyseverfahren durchgeführt.

1) Website-Besucher verstehen

Zunächst wurden mit einer explorativen Datenanalyse die Website-Besucher:innen und deren Bedürfnisse untersucht. Über die meisten der Besucher:innen lagen bereits Daten vor, da sie in der Vergangenheit bereits Käufe auf der Website getätigt hatten und dafür ein Konto angelegt hatten. Darüber hinaus wurde untersucht, über welche Kanäle die Besucher:innen in den Online-Shop gelangten, beispielsweise über Google oder Facebook. Informationen zu Gerät, Standort, Browser und Betriebssystem waren ebenfalls verfügbar.

Anhand dieser unterschiedlichen Parameter wurden die Benutzerdaten einem Analyseverfahren, dem sog. Clustering, unterzogen, bei dem die Website-Besucher:innen aufgrund ihrer Ähnlichkeiten in verschiedenen Eigenschaften in Gruppen („Cluster“) eingeteilt wurden.

Beispiel: Besucher über Android-Smartphones und Chrome-Browser, die zwischen 17 und 19 Uhr am Samstag auf der Website sind, kaufen eher familienbezogene Produkte.

Daraufhin konnte man neue Website-Besucher:innen aufgrund der verschiedenen Eigenschaften meist recht eindeutig einem Cluster zuordnen, da ähnliche Besucher:innen tendenziell ein ähnliches Verhalten auf einer Website zeigen. Dieses Clustering lieferte dem Unternehmen bereits wertvolle Informationen. So konnten auf dieser Informationsgrundlage individuelle Marketingstrategien für verschiedene Zielgruppen entwickelt, das Werbe-Targeting angepasst und spezifische Sonderangebote erstellt werden.

Beispiel: Besucher über Android-Smartphones und Chrome-Browsern, die zwischen 17 und 19 Uhr am Samstag auf der Website sind, bekommen ein Sonderangebot für ein familienbezogenes Produkt, wie beispielsweise ein Babyfon ausgespielt.

In vielen Fällen reicht eine solche Analyse bereits aus, um die Conversion-Rate eines Online-Shops spürbar zu steigern. In diesem Projekt wurden jedoch noch zwei weitere Analyseschritte durchgeführt.

2) Conversion Path verstehen und Engpässe nachvollziehen

Der nächste Schritt bestand darin, den Conversion Path der Kund:innen zu untersuchen. Der Conversion Path umfasst alle Handlungen von Kund:innen vom Ankommen auf der Website über den Besuch verschiedener Seiten bis hin zum finalen Kauf bzw. Kaufabbruch. Bei der Analyse wurden alle Conversion Paths auf Gemeinsamkeiten und Unterschiede untersucht, um bestimmte Muster abzuleiten. Von besonderem Interesse waren mögliche Gründe, aus denen Besucher:innen ihre Sitzung vor Kaufabschluss abbrachen. Es stellte sich heraus, dass Besucher:innen ihre Sitzung vor allem dann abbrachen, wenn es für ein Produkt kein Produktvideo gab bzw. das Produktvideo nicht gefunden wurde. Diese mangelnde Produktinformation konnte anschließend gezielt bearbeitet werden, woraufhin sich die Conversion-Rate deutlich verbesserte.

3) Next-best-Action vorhersagen

Im dritten Schritt des Projektes zur Steigerung der Conversion-Rate wurde der Ansatz der Next-best-Action (NBA) gewählt. Damit wurde hier ein weiterer Schritt von der reinen Analyse von bereits vorhandenen Daten hin zur Vorhersage zukünftigen Verhaltens gewählt.

Was bedeutet Next-best-action? Next-best-action (NBA) ist eine Marketingstrategie, die darauf abzielt, Informationen über einzelne Kund:innen zu sammeln und zu nutzen, um einen Kauf anzuregen. Wie der Name schon sagt, wird versucht zu ermitteln, welcher der nächste beste Schritt im Verkaufsprozess für jede:n einzelne:n Kunde:in ist.

Mithilfe der allgemeinen Informationen über die Website-Besucher:innen und der Conversion Paths konnten unterschiedliche Aktionen identifiziert werden, die einen Kauf wahrscheinlicher machen würden. Dazu gehörte z. B., den Besucher:innen das Produktvideo anzuzeigen, einen Rabatt-Code oder ein Sonderangebot für eine spezielle Produktkategorie anzubieten oder ein Chat-Fenster für den Kundensupport zu öffnen.

Somit half die NBA-Vorhersage dabei, die Conversion erneut deutlich zu steigern, indem für jede:n Website-Besucher:in eine individuelle Aktion vorgeschlagen werden konnte.

Ergebnisse:

In diesem Projekt konnte die Marketingabteilung des Elektrogeräte-Herstellers durch drei verschiedene Analyseansätze die Conversion-Rate im Online-Shop deutlich verbessern:

  • Mithilfe des Clustering war das Unternehmen in der Lage, individuelle Marketingstrategien für verschiedene Zielgruppen zu entwickeln, das Werbe-Targeting anzupassen und spezifische Sonderangebote zu erstellen.
  • Durch die Analyse der Conversion Paths konnten produkt- und produktbeschreibungsspezifische Engpässe identifiziert und anschließend gezielt behoben werden.
  • Mit der NBA-Analyse konnten nächste beste Schritte für jede:n einzelne:n Kunde:in bestimmt und automatisch ausgelöst werden.

How to tackle lack of data: an overview on transfer learning

1, Data is the new oil, but labeled data might be closer to it

Even though we have been in the 3rd AI boom and machine learning is showing concrete effectiveness at a commercial level, after the first two AI booms we are facing a problem: lack of labeled data or data themselves. The increasing number of papers on deep learning demonstrate that researches on AI have developed rapidly recently. If architectures of neural networks and supervised learning are all you know about deep learning, you will be overwhelmed by complications of topics studied these days, for example generative models, making more compact neural net models by for example knowledge distillation, and explainable AI (XAI). Those researches are often conducted on easily available benchmark datasets which you can easily download, often with corresponding ground truth data (label data) necessary for training. However once you try to apply the techniques to more specific data, you usually cannot prepare enough label data which theoretical researches assume. Thus among fascinating deep learning topics, in this article I am going to pick up how to tackle lack of label or data themselves, and transfer learning. Transfer learning is a technique of machine learning to take advantages of knowledge learned in one dataset to deal with a task in another dataset. Presumably due to this fact, Andrew Ng, in his presentation in NeurIPS 2016, gave a rough and abstract predictions of how transfer learning in machine learning would make commercial success like white lines in the figure below. The explanation is straightforward, and given the trends in topics of researches on machine learning these days, this prediction is actually right. But at the same time, in my opinion supervised learning, transfer learning, and unsupervised learning cannot be clearly separated like the graph originally suggested by Andrew Ng. Those fields complement each other, and one can easily shift to another.

Source: https://ruder.io/transfer-learning/ The lines and texts in white are based on explanations by Andrew Ng. The orange cells are placed at random, so not that they represent commercial success of each field.

Along with the rapid progress of deep learning mentioned above, a lot of hypes and catchphrases regarding big data and machine learning were made, and an interesting one is “Data is the new oil.” That might have been said only because big data is sources of various industries. But I would say, the characteristic is more striking in training data for machine learning. Distributions of training data for machine learning are more complicated like various energy resources besides oil in the world. Labeled data might be also like uranium. Just as uranium-235 accounting for only less than one percent of uranium in the world can be used to generate energy, only a part of massive data in the world is labeled such that they can be used for supervised machine learning. And as uranium-235 is used effectively jointly with less active uranium-238, labeled data show greater potentials with unlabeled data. And training data for machine learning have another unpleasant analogy to energy resources. Like most mainstream energy resources, only limited companies or institutions would be able to mine and refine huge labeled datasets with gigantic computation resources, and most people more or less need to rely on that for their business. Even though alternative renewable energy resources are proposed, principal energy resources are indispensable for making industries stable. As well, even though a lot of techniques actually have been proposed to lack of data, it often turns out just fine-tuning pre-trained models is the most practical, which need huge datasets and rich computational resources. And I think recent success in for example BERT or GPT made this trend more visible.

*I am sorry in a case I am mistaken about energy resources. I just wanted to come up with some cool metaphors.

But I still think knowing about transfer learning more comprehensively would be effective. That is partly because I have been working on relatively unique data which are hard to even label. As I was studying computer vision (CV) in plant science field, I frequently saw relatively unique data obtained with special apparatuses. Such data are for the most part look far from very general dataset, which huge pre-trained models are trained on. At the same time such plant data have very complicated structures and hard to label. And also in my work, have to detect certain values in various formats in very specific documents, in German. Such data are far from general datasets, and even labeling is hard in that case. We have to carefully tackle lack of data every time on each type of data in that case.

In this article I would first like to explain in the first place what it is like to lack data and next introduce representative techniques to tackle lack of labeled data. Many of them are classified to transfer learning, but other techniques like unsupervised learning or self-supervised learning are used in them or share a lot in their ideas. Thus my main purpose of writing this article is to let you have a richer view on transfer learning. And you would see “transfer learning” these days are mainly about fine-tuning of pre-trained models. Also how to tackle lack of data or labels is in other words how to efficiently achieve good performance in machine learning. Thus even if tons of high quality labeled data are at your disposal, learning those ideas would be still effective to you. I hope you could find some hints of machine learning through my articles.

2, What does lack of data or labels mean in the first place?

We need to first consider what lack of labels or data means, and my answer to the title of this section is “It depends.” The more data you have, the better performances you get. And the bigger machine learning models are, the more data they usually need for training. I assume that people reading this article more or less understand neural networks and how they are trained with back propagation. But let’s review the process here. Most machine learning frameworks are more or less expressed like the figure below unless reinforcement learning is considered. The ultimate purpose of machine learning is to train a model f(\boldsymbol{x}_n;\boldsymbol{\theta}) by adjusting parameters \boldsymbol{\theta}. And the parameters \boldsymbol{\theta} are optimized so that a loss function L is minimized. If it is a supervised learning, the a value of a loss function is denoted L(f(\boldsymbol{x}_n, \boldsymbol{\theta}), \boldsymbol{y}_n) =L(\hat{\boldsymbol{y}}_n, \boldsymbol{y}_n), and it gets smaller as f(\boldsymbol{x}_n, \boldsymbol{\theta}) gets closer to \boldsymbol{y}_n. That is, \boldsymbol{y}_n is giving supervision to adjust f(\boldsymbol{\theta}) via L(\hat{\boldsymbol{y}}_n, \boldsymbol{y}_n). And in a case of unsupervised learning, a loss function is L(\hat{\boldsymbol{y}}_n), which is often heuristically handcrafted.

The very first problem from lacking training data you would learn is overfitting. That is, a machine learning model can be specialized too much for a training dataset, and it loses generalization to other data from the same dataset. It is like students with little imaginations and flexibility gradually memorizing all the answers in a textbook and failing to answer new questions they have not encountered yet. Overfitting is judged by relations of training and validation loss like in the graph below. Training loss in blue indicates how the students adjust to the textbook. The smaller the training loss is, the more they memorizes from the textbook and the less flexible they are. The orange line indicates their performance in newly appeared questions in tests. The smaller the validation loss is, the better the students perform on tests. Thus the students should stop learning with the textbook when the validation loss is about to increase. This is called early stopping in machine learning. And if you increase training data, the orange graph usually shifts to the right side, usually providing smaller validation loss, namely better performance. An important point is, this ideal relations of training and validation losses will not appear if sizes or expressivity of a model is not enough. Thus the more training data you use, the more parameters you need for the model to enhance its expressivity.

 

*Depending on sizes of training data, the curve of training loss also changes, so please bear it in mind that this graph is not correct and is very simplified.

What I said so far might sound too elementary. My point is, the more data you have, and the bigger computation resource you have, the better performance you get. In other words, machine learning has scalability with data and parameters. This characteristic is clearly observed in models in natural language processing (NLP) and computer vision (CV) like in the graphs below. When I read some papers,often I am very fascinated by their performances. But sometimes it turns out that the methods are mainly creatively in terms of how they increase training data, which is personally boring. And even if performance of GPT looks astonishing, I cannot really like them because of this simple fact.

However another important point is, conversely you don’t need to increase training data or parameters of a model once it achieves an ideal score in metrics. When you make a toy model with small training data, as long as your clients or co-researchers are already happy, that is enough. Therefore lack of data or labels has to be discussed depending on sizes of machine learning and their performances you expect. Given those points mentioned so far, my answer to the question “What does lack of data or labels mean?” would rephrased like “If your model is properly designed to reach the performance you expect and it starts overfitting, you are facing lack of data.” And such decisions basically has to be made based on experiments.

3, Types of lack of data

Even though I explained lack of labels or data is a contextual matter, the problems actually exist at any case. That is, you often fail to achieve ideas accuracy partly due to lack of training data. I would like to classify types of situations of data of label shortage as below.

We should first think about the case where lack of labels does not matter in the first place. If you can analyze data with statistical knowledge or unsupervised machine learning, just extracting data without labeling would be enough. And sometimes ad hoc analysis with simple data visualization will help your decision makings. And some dashboards made from those unlabeled data will already give you some insights into data.

The next case is that, popular machine learning fields with enough investments usually have huge datasets that huge academic institutes or companies have been preparing.  For example KITTI dataset, which include labels like trajectories and depth data, is by Karlsruhe Institute of Technology and Toyota Technological Institute. Such datasets are useful for self-driving-related researches, and many types of ground truth data are provided such as odometry, depth, opticla flow, detection. This kind of data might be considered “enough” only because they are enough for training machine learning models and quantitatively evaluating them in papers, regardless of practical usefulness at a commercial level. But at any rate, popular fields with large benchmark datasets are likely to get investments for commercial uses.

Next let’s see cases of data shortage. You should also keep it in mind that there are also several types of situations of data shortage. In fact there are cases where certain labels are supposed to be scarce such as classifications of imbalanced data, for example anomaly detection, judging spam mails,  or medical examination. In those problems only some percent of data are classified as “errors,” “spam,” or “disease,” and others are classified as “normal.” Just keeping classifying data into “normal” would give maybe more than 95% accuracy. But finding the rest some percent accurately is much more important. In this case model performances need to be evaluated with ROC curves, namely relations of true positives and false positives.

The next type is more related to cases assumed in transfer learning. Some data are in the first place very expensive to obtain. For example CT images have to be stored by special medical apparatuses as you know. And even if a lot of CT images are already obtained, annotating the images often needs professional skills, thus its annotations cost is high. Another case of high annotation cost is for example detection or segmentation of objects in images. Even if you can collect numerous images on the Internet, annotating bounding boxes or pixel-wise segments require a lot of time. Annotating around 1000 images  for classification might be ok, but annotating them at a pixel level is really time consuming. If you have a tablet, I would like you to paint each segment of objects in a picture with different colors. And you should multiply the time spent by 80,000, as many as the training images needed for Mask R-CNN, a popular model for instance segmentation. As you can imagine, it is a huge tediou work. Even preparing some 50 labeled images for fine-tuning is paiful, and even annotations for computer vision tasks itself is also a field of deep learning.

*I would say medical image processing is a relatively popular field in CV with deep learning, and there are several famous datasets on this field.

4, An overview on ways for dealing with lack of labeled data

I am going to first roughly introduce what kind of approaches can be taken to deal with lack of labeled data or data itself, but you should also keep it in mind that they are not clearly separated. Just as I am going to explain, one type of techniques can easily shift to another type. You should flexibly switch among them depending on your situations. And also please keep it in mind that these are well-studied areas, and tons of ingenious papers are announced one after another, usually giving slight changes in their performances. Problems I point out about each technique might not be a problem anymore with recently published researches on researches currently peer-read. It is hard to prove that something does not exist. Given those points, I think it is convenient to classify technique of dealing with label or data shortage as below.

Through this article, ideas of domains are important. A domain simply means a combination of a dataset and a task with it. Transfer learning is a family of machine learning techniques to make uses of knowledge learned in a domain to another domain, and the former is called a source domain, the latter a target domain. And discrepancies between a source domain and a target domain is called a domain shift. The figure below abstractly visualize examples of domains and domain shifts. Intuitively it is easy to imagine that face a CV task and an NLP task have bigger domain shifts than domains of leaf images taken from different angles, but quantitatively evaluating domain shifts is in practice hard, and I am not going to introduce the topic because that will need a lot of mathematics.

Instead of formulating transfer learning, I would like to take learning languages as an intuitive example of transfer learning. Most people master at least one native language before learning another one. Baby brains are a kind of fantastic machine learning models, and after overcoming many obstacles they master native languages. And people take advantages of their mother tongues to learn another language. Usually they learn foreign languages by comparing structures of translated sentences. And naturally, if both a foreign language and your language have analogies like grammatical cases or genders in common, language learning would be easy. In other words, proficiency in one language is helpful in leaning some language. But it is also possible that your native language badly affects learning the second language, due to grammatical structures, pronunciations. The case of a source domain deteriorating performances in a target domain is called negative transfer and contexts of transfer learning.

*I know similarities languages are not the sole and definite barometers of effectiveness in learning foreign languages. Sizes of economy or markets in a country would also affects English language acquisition of people there. But at least it is unfair to compare for example German or Dutch people learning English with Japanese, Chinese people learning it. Unlike Eastern Asian people who have to learn thousands of characters to at least read decent texts or who use very different grammars, European people obviously can use “transfer learning” to learn English.

5, Increasing training data

When you lack data or labels, the most straightforward and often quick solution is to just increase data. The two topics I will cover in this section are mainly conducted in one domain.

Data augmentation

Data augmentation is one of the first techniques you would learn to mitigate overfitting of machine learning, which is in short caused by lack of data. The idea is very simple and it is implemented well in deep learning libraries, so I would only briefly talk about it here. The idea of data augmentation is simply transforming input data by for example flipping, rotating, zooming, changing colors. By doing so for example an input image \boldsymbol{x}_n of a butterfly below with a label of \boldsymbol{y}_n = \text{Butterfly} can be converted to more than 6 images. This corresponds to getting a converted \boldsymbol{x}'_n= g(\boldsymbol{x}_n) in the machine learning outline in the last section. And this process is the same as increasing the size of a dataet \mathcal {D}. And one point you have to be careful is, you must not change \boldsymbol{x}_n too much to change corresponding \boldsymbol{y}_n. For example if \boldsymbol{x}_n is distorted too much, it cannot be recognized as \boldsymbol{y}_n anymore even by humans. Or if you rotate an image of a digit 6 180 degrees, its becomes 9. Recent researches focus on automatically find what kind of data augmentation is effective by using for example reinforcement learning.

Here let me take an example of data augmentation technique that would be contrary to your intuition. A technique named mixup literally mix up data with different classes and their labels. In classification problems, labels are expressed as one-hot vectors, that is only an element corresponding to a correct element is 1 and the others are 0. In a case of binary dog-or-cat classification, each label is \boldsymbol{y}_n = (1, 0)^T or \boldsymbol{y}_n = (0, 1)^T, respectively. In data augmentation, distorting data too much is a taboo because label data is contaminated, but in mixup you literally mix up labels. Randomly choosing a two inputs \boldsymbol{x}_n , \boldsymbol{x}_{n'} and a  number \lambda \in [0,1], you prepare a input and label pair (\lambda \boldsymbol{x}_n + (1 - \lambda) \boldsymbol{x}_{n'},  \lambda \boldsymbol{y}_n + (1 - \lambda) \boldsymbol{y}_{n'}). The figure below is an example of a mixing up a cat input and a dog input, and corresponding labels. It is known augmenting training data like this improves classification performances. It is said this is partly due to machine learning models effectively learning decision boundaries. In classification ambiguous inputs are bottlenecks, so learning to giving ambiguous outputs to ambiguous inputs can enhance classification abilities.

*One-hot-encoded labels are called hard labels, and otherwise soft labels. Recent topics in deep learning, such as lottery hypothesis, knowledge distillation, imply that whether supervising labels are hard or not is important in deep learning. Hopefully I would like to explain why little by little in my articles.

6, Active learning

Active learning is about how to annotate data and get labeled data efficiently. Labels of data do not equally contribute to enhancing machine learning models, and labels actually have qualities. Even if you give apparently similar images with the same label to machine learning models during training, the models cannot learn so much from the pair of data. You need to efficiently dig data to know its distribution by giving labels to samples. I think a good metaphor is geological survey by excavating with some boring. In order to know substances or features of ground, some earth need to be sampled with boring. But you cannot freely penetrate everywhere mainly due to costs. They need to be sampled one by one due to uncertainty about the ground.

 

Similar approaches are often taken in machine learning or statistics, that is estimating distributions of data with a small size of samples is an important idea. A basic idea for doing that is you sample or annotate data which decreases uncertainty of your model the most. The figure simply exhibits the idea. We want to regress a data distribution with the red curve, and the cross marks can be sampled from the distribution. And the part filled with light blue shows uncertainty of the model to predict a value of y for a x. When you want to regress the data with as few samples as possible, data points should be sampled from the parts with great uncertainties. And by doing so, you can see that the data is regressed efficiently with few samples.

We have seen that modeling uncertainty is the key to active learning, and that can be applied to annotations of data in deep learning. An example of the process is displayed below, and in this case a deep neural network model (DNN model) is trained with some labeled data, and you give some signals for data annotations based on uncertainty of outputs of DNN models. And human annotators prioritize giving labels to the data. Such uncertainly can be estimated by using entropy of outputs or modeling data distributions.

 

But when you get a certain amount of labels, the situation will be the same as semi-supervised learning, which I will explain next. That is, you might be already able to make the most of the labels so far with the help of unlabeled data. You should consider stopping labeling and start labeling depending on situations. And importantly, starting naively annotating data might become a quick solution rather than thinking about how to make uses of limited labels if extracting data itself is easy and does not cost so much. “Shut up and annotate!” could be often the best practice in practice. And annotations would be an effective way for exploratory data analysis (EDA), so I recommend you to immediately start annotating about 10 random samples at any rate.

7, Dealing with lack of labels in a single domain

In many cases, data themselves are easily available, and only annotations costs matter. The following two topics consider such cases, and again only one domain is considered. But by the end of this article you would see that other techniques covered in this article have a lot of analogies with topics introduced here.

Semi-supervised learning

Semi-supervised learning is a type of supervised learning where only limited labels are available in one domain. This is important in because many of other techniques in this article can be seen as semi-supervised learning from certain points of views. The figure below shows an intuition on semi-supervised learning in a case of classification task. In this case, original data distribution have two clusters of circles and triangles and a clear border can be drawn between them. But only with limited labeled data, decision boundaries would be ambiguous. However in fact, with a help of unlabeled data in dotted lines, machine learning model might be able to recognize two clusters with a help of unlabeled data. In other words, unlabeled data help models learn distribution of data. this might be natural as clusters of data can be estimated with unsupervised learning.

*As I have already mentioned, active learning could soon shift to semi-supervised learning, and it might be worth trying it before finishing labeling. But suspending labeling and resuming it later might not be efficient. At any rate you need to be flexible depending on situations.

Semi-supervised learning is applicable to several tasks, not only classification. I explained that normal supervised learning is adjusting parameters \boldsymbol{\theta} of a model f(\boldsymbol{\theta}) so that it minimize loss function L(\boldsymbol{\theta}, \mathcal{D}_{\text{L}}) for a labeled dataset \mathcal{D}_{\text{L}}. In semi-supervised learning, we assume that usually a bigger unsupervised dataset \mathcal{D}_{\text{UL}} is available in the same domain. And semi-supervised learning optimize \boldsymbol{\theta} by jointly minimizing L(\boldsymbol{\theta}, \mathcal{D}_{\text{L}}) + L'(\boldsymbol{\theta}, \mathcal{D}_{\text{UL}}) after designing a loss function L'(\boldsymbol{\theta}, \mathcal{D}_{\text{UL}}) for the unlabeled dataset. There are following 3 major ways of semi-supervised learning depending on how you design a L'(\boldsymbol{\theta}, \mathcal{D}_{\text{UL}}).

  • Consistency regularization: adding slight changes to data \boldsymbol{x}_{\text{UL}} in \mathcal{D}_{\text{UL}} and get \boldsymbol{x}'_{\text{UL}}. And training f(\boldsymbol{\theta}) so that f(\boldsymbol{\theta}, \boldsymbol{x}_{\text{UL}}) and f(\boldsymbol{\theta}, \boldsymbol{x}'_{\text{UL}}) give out a consistent output.
  • Pseudo label: after training f(\boldsymbol{\theta}) with \mathcal{D}_{\text{L}}, using some estimations f(\boldsymbol{\theta}, \boldsymbol{x}_{\text{UL}}) as labels of \mathcal{D}_{\text{UL}} .
  • Entropy minimization: encouraging outputs f(\boldsymbol{\theta}, \boldsymbol{x}_{\text{UL}}) to have smaller entropy.

More or less similar ideas show up in different transfer learning techniques, so it would be effective to learn the three semi-supervised learning ideas above.

Self-supervised learning

Self-supervised learning is often counted as unsupervised learning. Both unsupervised and self-supervised learning do not need label data, but especially when labels generated by processing themselves, that is often called self-supervised learning. A representative case of using self-supervised learning is auto-encoder. Simpler labels can be generated from input data themselves with elementary data processing. For example in a case of image processing, by rotating an input image 0, 90, 180, 270 degrees respectively, a classification task of estimating rotation degrees can be made. Another case is estimating the original input image after some simple image processing (for example colorization).  These simple tasks generated solely from an input is called pretext task. And in a case of image processing, deep learning models can be prompted to learn image features .

Source: https://atcold.github.io/pytorch-Deep-Learning/en/week10/10-1/

Pretext tasks are applicable also to other fields for example NLP. A very simple task is hiding a part of an input sentence, and let neural networks estimate the blank word. And this is a basic idea of how to train BERTs, famous pre-trained NLP models. BERT models are trained this way with a huge and very general corpus without any specific topics. By doing so BERT model can already learn to detect some clusters of meanings in texts, as I visualize in the next section. But if you fine-tune BERT models with labeled texts with very specific topics, that often fails to achieve satisfying performance. In that case, the BERT models have to “get used to” the new dataset. In that case, BERT can “get used to” the new dataset by applying self-supervised learning on the new dataset. This tutorial of Huggingface demonstrates this with an example of adjusting a BERT model trained with Wikipedia to the IMDb dataset.

In the case above, the BERT model is fine-tuned with relatively lots of unlabeled data and after that trained with fewer labels. As a whole this can be seen as semi-supervised learning ,with fewer labels of the IMBb dataset and more unlabeled data. Also the ideas of pretext tasks, which prompt models to give consistent outputs given preprocessed inputs, have some analogies with consistency regularization in semi-supervised learning.

*The Huggingface tutorial says, they fine-tune a pre-trained BERT model trained in a self-supervised way to adjus it, and they call it “domain adaptation.” As you can see from the statement, distinctions of topics covered in this article can be just ambiguous.

8, Dealing with lack of data or labels over several domains

Another approach for tackling label or data shortage is taking advantages of other domains, which are usually larger and have enough labels. And such techniques is called transfer learning as I mentioned. It seems like transfer learning in business refers to “fine-tuning” explained below, but in academic contexts it is often also said transfer learning is almost synonym to “domain adaptation.” At any rate, my point is it would be more important to have comprehensive view on the techniques rather than clearly distinguishing them.

Fine tuning

Fine tuning would be the easiest way of transfer learning, and at the same time it is very powerful. Even though I am going to introduce other technique of transfer learning, more often than not it turns out that fine tuning can compensate them. Here I will only explain what it is like to use fine-tuning. I would say using fine-tuning is easy like using instant coffee. Conventionally you needed to train your original model with your own data, and that is very affected by sizes of data you have. I would say, that was like making coffee or coffee cakes from coffee you made from beans. But by using pre-trained models already trained somewhere with huge datasets, you can use models which can already more or less recognize data. The idea was very normal already in the field of CV, and NLP got the same idea with the advent of BERT, or already with word embeddings. That is like people learned to use instant coffee instead of roasting and brewing coffee every time.

How such instant coffee is made depends on which type of deep learning is used on a huge dataset. Backbone CNN is usually trained on ImageNet dataset with supervised learning of a classification task. In case of BERT, it is trained with a huge corpus with a pretext task of estimating blank words of input sentences, which is classified to self-supervised learning. Let me more practically what the “coffee syrup” means. Machine learning is at any rate just mapping of tensors or vectors. In CV, an input images as a tensor is converted into a a vector or a tensor, and tasks like image classification are conducted with the converted tensor or vector. In case of an NLP task, usually a sequence of vectors is converted to a vector or another sequence of vectors. And these resulting tensors of vectors from models are the very “coffee syrup” I am talking about. An important point is, fine-tuning also considers transfer learning between different tasks. Backbone CNNs are usually trained with classification, BERT with self-supervised learning, but the there are a variety of final tasks. They are called downstream tasks. In other words, you don’t necessarily drink instant coffee as coffee.

 

The two figures below are visualizations what the “instant coffee syrup” means. I processed random N images in a dataset with a pre-trained backbone CNN, and I got corresponding D dimensional vectors, that is a N\times D tensor. And I applied t-SNE to reduce its dimension from D to 2 and got a N\times 2 tensor.  The figure below shows arrangements of input images in the 2 dimensional space. As you can see, semantically similar images get closer.

Just as well, if you process random texts with BERT and apply a dimension reduction, you get a visualization like below. As well as the figure above, texts in similar topic get closer.

To make it catchy I expressed them as “coffee syrup” but this is a kind of how so-called AI sees data. Images and texts are just vectors or tensors on computer, and AI process another set of tensors of vectors in spaces which make sense to them.

Fine-tuning is quite easy. You have only to train a pre-trained model you downloaded just like normal supervised learning with your dataset. And when you train CV models with backbone CNN, the backbone is almost automatically downloaded. You have to be careful about some points, for example you have to set learning rate smaller. Let me avoid too detailed points in this article. Hopefully in the future, I’d like to write about more practical fine-tuning tips.

Domain adaptation

Domain adaptation is another family of techniques to make uses of knowledge gained in one domain in another domain. Domain adaptation is a Domain adaptation is these days often used as almost a synonym of transfer learning. But papers on domain adaptation usually assume to handle the same tasks both in a source and a target domain. So I would say domain adaptation is a subfield of transfer learning. Domain adaptation is more of how to tackle deterioration of machine learning performances when trained models are applied in different domains. Based on how much labels are available in each domain, domain adaptation is classified to several types. And unsupervised domain adaptation (UDA), where labels are available only in a source domain, is considered as the most challenging and studied well.

*Another explanation I often hear about domain adaptation is, when a models trained on a dataset is trained on another data, domain adaptation can be used to mitigate decreases in performance. I think in this context, performance of the model on the source domain is not discussed. When you apply some retraining with a new dataset, performance of the model on the source domain often drastically decrease. This is called catastrophic forgetting, and techniques like continuous learning are studied to tackle this problem. I have not really seen continuous learning in contexts of domain adaptation, but I thin these are related.

There several approaches in domain adaptation, and one frequently used approach is using adversarial loss. As we saw with the example of getting “coffee syrup,” data is first mapped into a certain space, and this is often called feature extraction. And outputs with the feature extractor are processed are processed more to give task-specific results with some networks. Often in domain adaptation, we put a domain discriminator network right after the feature extractor. And the domain discriminator classifies whether the features extracted come from the source or target domain. The feature extractor tries to extract features the domain have in common, and the domain discriminator tries to distinguish them, and two networks compete. In this way, the feature extractor and the domain discriminator form generative adversarial network (GAN), and the feature extractor learns to extract features that are hard to distinguish their domains. Feature extractor is trained so that it extract domain invariant features, for example edges and silhouette.

As well as in other transfer learning techniques, one ultimate goal of UDA is training a deep learning model only with synthetic labeled data, for example CGI, and apply the model on a totally unlabeled dataset. Converting a source domain to look like a target domain with Cycle GAN is an often used approach in domain adaptation. In domain adaptation a source domain is supposed to be easier to annotate. The figure below is an example of converting a black and white cell images  to colored images.

*You could easily try converting data with Cycle GAN by preparing two datasets, and I made the converted data by myself. But you need at least one GPU to try that.

However some people insist that usefulness of UDA is very questionable. In the first place, if you do not have any labels on the target domain, that means you cannot evaluate anything qualitatively on the dataset of interest. And if you can prepare some of evaluation data or labels, applying other techniques like fine-tuning might be enough.

Meta learning and few-shot learning

One simple way to explain meta learning is that, it is a machine learning technique teach models to learn efficiently. We can also say that it is a transfer learning case where target domains are unknown.  A famous meta learning method is Model-Agnostic Meta-Learning (MAML). MAML is used to get an ideal parameter \boldsymbol{\theta} which can be quickly and effectively used to new tasks. Like in the figure below, \boldsymbol{\theta} reaches the generally convenient parameter shown as the black dot. And the parameter can quickly reach the parameters \theta_{i}^{\ast}, which effective for each task.

Another interesting application of meta learning is few-shot learning. Few-shot learning trains a classification model to learn to acquire classification ability based only on a very few samples. By letting the models learn classification tasks over many episodes, the model learn comes to learn efficiently from limited data samples at a test phase. The figure below shows a case of few-shot learning, where a model learns some episodes of 3-class classifications with only 4 samples per class. Few-shot learning attempts to enable human-level flexibility of perception. MAML is known to be effective also for few-shot learning.

However, studies these days do also show that fine tuning pre-trained models with a few sample data show competitive results to those by few-shot learning. Similar things can be said about large language models like GPT. Chat GPT or GPT-3/GPT-4 for example can be fine-tuned with small extra training samples, and the logic behind is different from meta learning. Fine-tuning pre-trained models rather might be closer to human learning. Humans can effectively learn new topics based on what they have experienced so far. Thus again here, fine-tuning models can be an easier and realistic solution.

I have explained an overview of machine learning techniques for handling lack of data, and as you might have noticed, fine-tuning models could be enough in many cases. I am not sure how much other transfer learning technique would be widely as useful as fine-tuning at a business level. At least, I hope this article would be a rough guideline for machine learning tasks with small sizes of data or labels. And if you have a chance to work on very unique data with very few labels, you wouldn’t be able to rely so much on only naive fine tuning of pre-trained models. In that case, you tasks have your own problem, and you would have to be careful about your EDA, data cleaning, and labeling. In that case you should consider some techniques introduced here. Hopefully someday I would like to write more detailed tutorials with each transfer learning technique. And I hope you would be able to apply a variety of transfer learning locally, not only relying on huge resources of gigantic entities.  And that would lead to a more secure future, I guess.

Web Scraping mit und ohne Code

Wenn Sie ein kleines Start-up haben, möchten Sie möglicherweise nicht mit anderen Unternehmen zusammenarbeiten, um Marketinginformationen zu erhalten, die für Ihr Geschäft wichtig sind. Das ist aber nicht nötig, denn bei fast allen erfolgreichen Unternehmen bekommt man online alle Daten, die man braucht, nur mit eigenem Wissen und ein paar nützlichen Web-Tools.

Aber um die erhaltenen Daten in Ihren Marketing- und anderen Strategien verwenden zu können, müssen Sie sie in die richtige Form bringen. Wie kann man Daten aus dem Internet extrahieren, verarbeiten, bereinigen und auch visualisieren? Welche Aktivitäten sind legal und welche nicht?

Mit Web Scraping die richtigen Daten extrahieren

Web Scraping bezieht sich auf die Verwendung eines Programms oder Algorithmus, um große Datenmengen aus dem Internet zu extrahieren und zu verarbeiten. Egal, ob Sie Datenwissenschaftler, Ingenieur oder jemand sind, der große Mengen an Datensätzen analysiert, die Möglichkeit, Daten aus dem Internet zu sammeln, ist in allen Bereichen nützlich, vom Sport bis zum Vertrieb.

Angenommen, Sie haben Daten im Internet gefunden und es gibt keine direkte Möglichkeit, sie herunterzuladen. In diesem Fall wäre Web Scraping mit Code wie etwa mit R oder Python eine Gelegenheit, mit der Sie Daten in eine verwendbare Form extrahieren können, die importiert werden kann.

Parsen von Webseiten mit Beautiful Soup

Wenn Sie Jupyter Notebook verwenden, sollten Sie damit beginnen, die erforderlichen Module zu importieren – seaborn, pandas, numpy, matplotlib.pyplot. Wenn Sie Jupyter Notebook nicht installiert haben, können Sie es mit der online verfügbaren Anaconda Python-Distribution installieren. Stellen Sie sicher, dass sie die integrierte %matplotlib enthält, um Diagramme einfach anzuzeigen.

Um Webseiten zu analysieren, sollten Sie auch einige Bibliotheken importieren:

  • Das Modul urllib.request wird zum Öffnen von URLs verwendet.
  • Das Paket Beautiful Soup wird verwendet, um Daten aus HTML-Dateien zu extrahieren.
  • Die Beautiful Soup-Bibliothek heißt bs4, was für Beautiful Soup Version 4 steht.

Nach dem Importieren der erforderlichen Module sollten Sie die URL mit dem Datensatz angeben und an urlopen() übergeben, um den HTML-Code der Seite abzurufen:

url = “http://www.webseite-beispiel.com/beitrag”

html = urlopen(url)

Das Erhalten der HTML-Seite ist nur der erste Schritt. Der nächste Schritt besteht darin, ein Beautiful Soup-Objekt aus HTML zu erstellen. Dies geschieht, indem der HTML-Code an die Funktion BeautifulSoup() übergeben wird.

Das Paket Beautiful Soup wird zum Analysieren von HTML verwendet, d. h. es nimmt rohen HTML-Text und parst ihn in Python-Objekte. Das zweite Argument „lxml“ ist der HTML-Parser:

soup = BeautifulSoup(html, ‘lxml’)

type(soup)

bs4.BeautifulSoup

Mit dem Soup-Objekt können Sie interessante Informationen über die Website extrahieren, die Sie scrapen, z. B. den Titel der Seite abrufen:

title = soup.title

print(title)

Sie können auch den Text einer Webseite abrufen und schnell ausdrucken, um zu sehen, ob er Ihren Erwartungen entspricht. Sie können den HTML-Code einer Webseite anzeigen, indem Sie mit der rechten Maustaste auf eine beliebige Stelle auf der Webseite klicken und „Inspizieren“ („Inspect“) auswählen.

Als Nächstes können Sie die Suppenmethode find_all() verwenden, um nützliche HTML-Tags von der Webseite zu extrahieren. Beispiele für nützliche Tags sind <a> für Hyperlinks, <table> für Tabellen, <tr> für Tabellenzeilen, <th> für Tabellenköpfe und <td> für Tabellenzellen.

Wenn es Ihnen schwer fällt, die notwendigen Informationen mithilfe von Code aus der ausgewählten Website zu extrahieren, macht das nichts. Heutzutage gibt es mehrere wirklich hochwertige Software, mit der Sie dies in wenigen Minuten ohne Python-Kenntnisse tun können.

Auf diese Weise treffen Sie den neuen Standard im Web Scraping. Mit den besten Tools können Sie nicht nur Informationen aus allen öffentlichen Quellen extrahieren, sondern sie auch im gewünschten Format und am richtigen Ort speichern.

Ist Web Scraping legal?

Web Scraping gibt es schon lange und in guter Verfassung ist es eine wichtige Säule des Webs. „Gute Bots“ ermöglichen es beliebten Diensten und Unternehmen, viele nützliche Aktionen auszuführen:

  • Suchmaschinen, Webinhalte zu indexieren
  • Preisvergleichsdiensten, den Verbrauchern Geld zu sparen
  • Marktforschern, Stimmungen in sozialen Netzwerken einzuschätzen

Allerdings gibt es auch „Bad Bots“. Sie extrahieren Inhalte von einer Website mit der Absicht, sie für Zwecke zu verwenden, die außerhalb der Kontrolle des Eigentümers der Website liegen. Bad Bots machen 20 Prozent des gesamten Webverkehrs aus und werden verwendet, um eine Vielzahl von böswilligen Aktivitäten wie Denial-of-Service-Angriffe, Online-Betrug, Kontohijacking, Datendiebstahl, Diebstahl geistigen Eigentums, nicht autorisiertes Scannen auf Schwachstellen, Spam und Betrug mit digitaler Werbung durchzuführen.

Was ist also legal und was ist illegal? Web Scraping und Crawling ist per se nicht illegal. Schließlich kann man seine eigene Website problemlos scrapen oder crawlen.

Startups mögen es, weil es eine kostengünstige und leistungsstarke Möglichkeit ist, Daten zu sammeln, ohne dass eine Partnerschaft erforderlich ist. Großunternehmen nutzen Scraper zu ihrem eigenen Vorteil, wollen aber auch nicht, dass andere Bots in Bezug auf sie einsetzen.

Die meisten Websites haben keinen Web-Scraping-Schutz. Sollten Unternehmen Web Scraping verhindern? Während die Gerichte versuchen, über die Rechtmäßigkeit des Scrapings zu entscheiden, werden Unternehmen immer noch ihre Daten gestohlen und die Geschäftslogik ihrer Websites missbraucht.

Aber anstatt sich an das Gesetz zu wenden, um dieses technologische Problem letztendlich zu lösen, kann man es mit Anti-Bot- und Scraping-Technologien bekämpfen.

Interview Benjamin Aunkofer - Business Intelligence und Process Mining ohne Vendor-Lock-In

Interview – Business Intelligence und Process Mining ohne Vendor Lock-in!

Das Format Business Talk am Kudamm in Berlin führte ein Interview mit Benjamin Aunkofer zum Thema “Business Intelligence und Process Mining nachhaltig umsetzen”.

In dem Interview erklärt Benjamin Aunkofer, was gute Business Intelligence und Process Mining ausmacht und warum Unternehmen in jedem Fall daran arbeiten sollten, den gefürchteten Vendor Lock-In zu vermeiden, der gerade insbesondere bei Process Mining droht, jedoch leicht vermeidbar ist.

Nachfolgend das Interview auf Youtube sowie die schriftliche Form zum Nachlesen:


Interview – Process Mining, Business Intelligence und Vendor Lock

1 – Herr Aunkofer, wir wollen uns heute über Best Practice bei der Verarbeitung von Daten unterhalten. Welche Fehler sollten Unternehmen unbedingt vermeiden, wenn sie ihre Daten zur Modellierung aufbereiten?

Mittlerweile weiß ja bereits jeder Laie, dass die Datenaufbereitung und -Modellierung einen Großteil des Arbeitsaufwandes in der Datenanalyse einnehmen, sei es nun für Business Intelligence, also Reporting, oder für Process Mining. Für Data Science ja sowieso. Vor einen Jahrzehnt war es immer noch recht üblich, sich einfach ein BI Tool zu nehmen, sowas wie QlikView, Tableau oder PowerBI, mittlerweile gibt es ja noch einige mehr, und da direkt die Daten reinzuladen und dann halt loszulegen mit dem Aufbau der Reports.

Schon damals in Ansätzen, aber spätestens heute gilt es zu recht als Best Practise, die Datenanbindung an ein Data Warehouse zu machen und in diesem die Daten für die Reports aufzubereiten. Ein Data Warehouse ist eine oder eine Menge von Datenbanken.

Das hat den großen Vorteil, dass die Daten auf einer Ebene modelliert werden, für die es viele Experten gibt und die technologisch auch sehr mächtig ist, nicht auf ein Reporting Tool beschränkt ist.
Außerdem veraltet die Datenbanktechnologie nur sehr viel langsamer als die ganzen Tools, in denen Analysen stattfinden.

Im Process Mining sind ja nun noch viele Erstinitiativen aktiv und da kommen die Unternehmen nun erst so langsam auf den Trichter, dass so ein Data Warehouse hier ebenfalls sinnvoll ist. Und sie liegen damit natürlich vollkommen richtig.

2 – Warum ist es so wichtig einen Vendor Lock zu umgehen?

Na die ganze zuvor genannte Arbeit für die Datenaufbereitung möchte man keinesfalls in so einem Tool haben, das vor allem für die visuelle Analyse gemacht wird und viel schnelleren Entwicklungszyklen sowie einem spannenden Wettbewerb unterliegt. Sind die ganzen Anbindungen der Datenquellen, also z. B. dem ERP, CRM usw., sowie die Datenmodelle für BI oder Process Mining direkt an das Tool gebunden, dann fällt es schwer z. B. von PowerBI nach Tableau oder SuperSet zu wechseln, von Celonis nach Signavio oder welches Tool auch immer. Die Migrationsaufwände sind dann ein ziemlicher Showstopper.

Bei Datenbanken sind Migrationen auch nicht immer ein Spaß, die Aufwände jedoch absehbarer und vor allem besteht selten die Notwendigkeit dazu, die Datenbanktechnologie zu wechseln. Das ist quasi die neutrale Zone.

3 – Bei der Nutzung von Daten fallen oft die Begriffe „Process Mining“ und „Business Intelligence“. Was ist darunter zu verstehen und was sind die Unterschiede zwischen PM und BI?

Business Intelligence, oder BI, geht letztendlich um die zur Verfügungstellung von guten Reports für das Management bis hin zu jeden Mitarbeiter des Unternehmens, manchmal aber sogar bis zum Kunden oder Lieferanten, die in Unternehmensprozesse inkludiert werden sollen. BI ist gewissermaßen schon seit zwei Jahrzehnten ein Trend, entwickelt sich aber auch immer weiter, mit immer größeren Datenmengen, in Echtzeit usw.

Process Mining ist im Grunde eng mit der BI verwandt, man kann auch sagen, dass es ein BI für Prozessanalysen ist. Bei Process Mining nehmen wir uns die Log-Daten von operativen IT-Systemen vor, in denen Unternehmensprozesse erfasst sind. Vornehmlich ERP-Systeme, CRM-Systeme, Dokumentenmangement-Systeme usw.
Die Daten bereiten wir in sogenannte Event Logs, also Prozessprotokolle, auf und laden sie dann ein eines der vielen Process Mining Tools, egal in welches. In diesen Tools kann man dann Prozess wirklich visuell betrachten, filtern und analysieren, rekonstruiert aus den Daten, spiegeln sie die tatsächlichen operativen Vorgänge wieder.

Auch bei Process Mining tut sich gerade viel, Machine Learning hält Einzug ins Process Mining, Prozesse können immer granularer analysiert werden, auch unstrukturierte Daten können unter Einsatz von AI mit in die Analyse einbezogen werden usw.
Der Markt bereinigt sich übrigens auch dadurch, dass Tool für Tool von größeren Software-Häusern aufgekauft werden. Also der Tool-Markt ist gerade ganz krass im Wandel und das wird die nächsten Jahre auch so bleiben.

4 – Wie ist denn die Best Practice bei der Speicherung, Aufbereitung und Modellierung von Daten?

BI und Process Mining sind eigentlich eher Methoden der Datenanalytik als einfach nur Tools. Es ist ein komplexes System. Ganz klar hierfür ist der Aufbau eines Data Warehouses, dass aus Datensicht quasi so eine Art Middleware ist und Daten zentral allen Tools bereitstellt. Viele Unternehmen haben ja um einiges mehr als nur ein Tool im Haus, die kann man dann auch alle weiterhin nutzen.

Was gerade zum Trend wird, ist der Aufbau eines Data Lakehouses. Ein Lakehouse inkludiert auch clevere Art und Weise auch einen Data Lake.

Den Unterschied kann man sich wie folgt vorstellen: Ein Data Warehouse ist wie das Regel zu Hause mit den Ordnern zum Abheften aller wichtigen Dokumente, geordnet nach … Ordner, Rubrik, Sortierung nach Datum oder alphabetisch. Allerdings macht es auch große Mühe, diese Struktur zu verwalten, alles ordentlich abzuheften und sich überhaupt erstmal eine Logik dafür zu erarbeiten. Ein Data Lake ist dann sowas wie die eine böse Schublade, die man eigentlich gar nicht haben möchte, aber in die man dann alle Briefe, Dokumente usw. reinwirft, bei denen man nicht weiß, ob man diese noch braucht. Die Inhalte des Data Lakes sind bestenfalls etwas vorsortiert, aber eigentlich hofft man ja nicht, da wieder irgendwas drin wiederfinden zu müssen.

5 – Sie haben ja einen guten Marktüberblick: Wie gut sind deutsche Unternehmen in diesen Bereichen aufgestellt?

Grundsätzlich schon mal gar nicht so schlecht, wie oft propagiert wird. In beinahe jedem deutschen Unternehmen existiert mittlerweile ein Data Warehouse sowie Initiativen zur Einführung von BI, Process Mining und Data Science bzw. KI, in Konzernen natürlich stets mehrere. Was ich oft vermisse, ist so eine gesamtheitliche Sicht auf die Dinge, es gibt ja viele Nischenexperten, die sich auf eines dieser Themen stürzen, es aber nicht in Verbindung zu den anderen Themen betrachten. Z. B. steht auch KI nicht für sich alleine, sondern kann sowohl der Business Intelligence als auch Process Mining über den Querverweis befähigen, z. B. zur Berücksichtigung von unstrukturierten Daten, oder ausbauen mit Vorhersagen, z. B. Umsatz-Forecasts. Das ist alles eine Datenevolution, vom ersten Report von Unternehmenskennzahlen über die Analyse von Prozessen bis hin zu KI-getriebenen Vorhersagesystemen.

6 – Wo sehen Sie den größten Nachholbedarf?

Da mache ich es kurz: Unternehmen brauchen Datenstrategien und ein Big Picture, wie sie Daten richtig nutzen, dabei dann auch die unterschiedlichen Methoden der Nutzung dieser Daten richtig kombinieren.

Sehen Sie die zwei anderen Video-Interviews von Benjamin Aunkofer:

Interview Benjamin Aunkofer – Datenstrategien und Data Teams entwickeln!

 

 

 

 

 

 


 

Cloud Data Platform for Shopfloor Management

How Cloud Data Platforms improve Shopfloor Management

In the era of Industry 4.0, linking data from MES (Manufacturing Execution System) with that from ERP, CRM and PLM systems plays an important role in creating integrated monitoring and control of business processes.

ERP (Enterprise Resource Planning) systems contain information about finance, supplier management, human resources and other operational processes, while CRM (Customer Relationship Management) systems provide data about customer relationships, marketing and sales activities. PLM (Product Lifecycle Management) systems contain information about products, development, design and engineering.

By linking this data with the data from MES, companies can obtain a more complete picture of their business operations and thus achieve better monitoring and control of their business processes. Of central importance here are the OEE (Overall Equipment Effectiveness) KPIs that are so important in production, as well as the key figures from financial controlling, such as contribution margins. The fusion of data in a central platform enables smooth analysis to optimize processes and increase business efficiency in the world of Industry 4.0 using methods from business intelligence, process mining and data science. Companies also significantly increase their enterprise value with the linking of this data, thanks to the data and information transparency gained.

Cloud Data Platform for shopfloor management and data sources such like MES, ERP, PLM and machine data.

Cloud Data Platform for shopfloor management and data sources such like MES, ERP, PLM and machine data. Copyright by DATANOMIQ.

If the data sources are additionally expanded to include the machines of production and logistics, much more in-depth analyses for error detection and prevention as well as for optimizing the factory in its dynamic environment become possible. The machine sensor data can be monitored directly in real time via respective data pipelines (real-time stream analytics) or brought into an overall picture of aggregated key figures (reporting). The readers of this data are not only people, but also individual machines or entire production plants that can react to this data.

As a central data architecture there are dozens of analytical applications which can be fed with data:

OEE key figures for Shopfloor reporting
Process Mining (e.g. material flow analysis) for manufacturing and supply chain.
Detection of anomalies on the shopfloor or on individual machines.
Predictive maintenance for individual machines or entire production lines.

This solution scales completely automatically in terms of both performance and cost. It looks beyond individual problems since it offers universal and flexible scope for action. In other words, it will result in a “god mode” for the management being able to drill-down from a specific client project to insights into single machines involved into each project.

Are you interested in scalable data architectures for your shopfloor management? Or would you like to discuss a specific problem with us? Or maybe you are interested in an individual data strategy? Then get in touch with me! 🙂

Phishing: An Overlooked Threat to Business and Data

Data is the lifeblood of an organization and businesses that fail to embrace this fundamental concept risk losing future business opportunities or the company itself. The value of business and consumer data hasn’t been lost on cybercriminals, which is why phishing attacks and other data security threats have been more rampant through the years. As data technology becomes more advanced, reliance on data becomes a more significant opportunity for exploitation as even the most advanced data systems have their vulnerabilities. Businesses collect and process large amounts of data from several sources, making the protection of this data one of their main challenges.

The Threat to Business and Data

Today’s businesses demand a lot from the data they collect, and as such, also look for solutions that will help transform data into more tangible bits and pieces that will help promote business success. The role of AI in cybersecurity has been gaining mainstream attention because it helps automate the detection of security threats and other malicious activities within a company’s systems. 

Looking to more advanced solutions for cybersecurity is a prudent approach, and is highly recommended, especially for businesses that handle large amounts of data daily. Cybersecurity threats are no laughing matter, and they’re becoming more advanced and difficult to identify and address. The end goal remains the same, however—to gain unauthorized access to data and use it to harm a business or for personal gain. The threats come in many forms, including DDoS attacks, malware, and phishing attacks.

Cybersecurity threats are serious not only due to the potential loss of data but also because a data breach can cause irreparable harm to a business. Give cybercriminals access to sensitive data and there’s no telling what damage they’ll cause. Even a simple phishing scam can lead to a full-blown data breach, and these breaches rely on users making mistakes

Recognizing Phishing Scams

Before businesses can combat scams, they must train their employees on the detection and proper ways of addressing or preventing them altogether. Below are the most common types of phishing attacks done against businesses.

Phishing Emails

Phishing emails are common in a business email inbox—typically categorized together with spam emails. The difference is that they are not simply unsolicited marketing messages; they are designed to trick you into opening a malicious attachment or clicking on a link to a fraudulent website. Scammers often use an email address that resembles a legitimate business email address to confuse users into thinking that the email came from someone within the organization or a trusted partner or a third party.

Company Impersonation

This method is a type of phishing scam in which scammers try to impersonate your brand. This is often done via “domain spoofing” or using a fake but similar website or email domain designed to confuse the recipients of the email. It can be difficult to detect as a source of a data breach because it’s often unreported. Victims of these scams often aren’t aware that they’ve been duped until it’s too late.

Phone Phishing/Voice Phishing

This is similar to company impersonation but uses a different medium—Voice over Internet Protocol (VoIP). Most scammers who use this method go so far as to imitate the tone of voice and spiels of the brands they’re trying to impersonate. This is usually done in conjunction with other phishing scams to increase the chances of success.

Spear phishing

This is similar to phishing emails but with a more targeted approach. It involves more effort on the scammers’ part because it entails getting key information about a target. Using this method, the scammers send an email containing the name of a company officer and other personalizations to make the recipient believe that the email is legitimate. Unsuspecting users will be duped into sending money, pertinent information, or making payments to a fake vendor or partner. It’s a sophisticated scam that can often trick even tech-savvy individuals, depending on how skillfully the fake email message is crafted.

Email Account Takeover

It’s one thing to have scammers send fake emails and try to make them look legitimate, but it’s another when they get access to a legitimate email account and use it to get money from unsuspecting users, gain access to sensitive information, or cause harm to a business. It’s typically done as part of a cross-account takeover, in which a scammer gains access to a user’s email account, changes the password, and proceeds to make fund transfers from the compromised user’s bank and other financial accounts.

Why Cyberattackers Resort to Phishing Scams

Phishing scams are one of the most common cyberattacks that threaten the security of the company and personal data, with spear phishing being one of the main infection vectors. It’s a common scam because phishing is as easy; anyone can execute a phishing attack and be relatively successful with little to no investment. The flexibility of a phishing scam also makes it an infection vector of choice. With it, a scammer can steal sensitive data and gain access to user accounts.

The simplicity-to-value ratio is also a tempting draw of phishing. Despite how simple its execution is, it can help cybercriminals get their hands on large sums of money. $17,700 is lost every minute because of a phishing scam.

Lastly, users aren’t good at putting a stop to scams—even large companies face the threat of phishing today. Because of how sophisticated these phishing emails have become and the large number of emails people have to go through each day, it can distinguish a fake email from a real one. 

Fighting the Good Fight

Despite the evolving threats to data security, there are still ways you can fight them. Just like cybercriminals always find vulnerabilities to exploit, you can always find ways to mitigate or even counteract these measures. The key is in keeping an open eye and mind and keeping abreast of the available solutions that can help you keep your data and business secure. Protecting your digital assets often requires “digital measures.” It won’t hurt to let digital tools help you, but you and your organization should be at the heart of your data security. Educate yourself and your employees so your security isn’t only as good as the tool you choose to use.