Every data scientist has been there: a new data set and you’re going through nine circles of hell trying to build the best possible model. Which machine learning method will work best this time? What values should be used for the hyperparameters? Which features would best describe the data set? Which combination of all of these would lead to the best model? There is no single right answer to these questions because, as we know, it’s impossible to know a priori which method or features will perform best for any given data set. And that is where parameter optimization comes in.

Parameter optimization is an iterative search for the set of hyperparameters of a machine learning method that leads to the most successful model based on a user-defined optimization function.

Here, we introduce an advanced parameter optimization workflow that uses four common machine learning methods, individually optimizes their hyperparameters, and picks the best combination for the user. In the current implementation the choice of features and one hyperparameter per method are optimized. However, we encourage you to use this workflow as a starting point or a template if you have completely different data and customize it by including additional parameters into the optimization loop (and we will show where you could do that).

In this blog series we’ll be experimenting with the most interesting blends of data and tools. Whether it’s mixing traditional sources with modern data lakes, open-source devops on the cloud with protected internal legacy tools, SQL with noSQL, web-wisdom-of-the-crowd with in-house handwritten notes, or IoT sensor data with idle chatting, we’re curious to find out: will they blend? Want to find out what happens when IBM Watson meets Google News, Hadoop Hive meets Excel, R meets Python, or MS Word meets MongoDB?

Follow us here and send us your ideas for the next data blending challenge you’d like to see at willtheyblend@knime.com.

Today: The GDPR Force Meets Customer Intelligence – Is It The Dark Side?

Authors: Rosaria Silipo and Phil Winters

The Challenge

The European Union is introducing the General Data Protection Regulation on May 25 2018. This new law has been called the world’s strongest and most far-reaching law aimed at strengthening citizens' fundamental rights in the digital age. The new law applies to any organization processing personal data about a citizen or resident in the EU regardless of that organizations’ location. It is definitely a force to be reckoned with as the fines for non-compliance are extremely high.

Many of the law’s Articles deal with how personal data are collected and processed and include a special emphasis, guidelines and restrictions on what the EU calls “automatic profiling with personal data”. So what has that to do with us data scientists? Many of us use methods and machine learning to create new fact-based insight from the customer or personal data that we collect so that we can take decisions, possibly automatically. So the law applies to us. But that Customer Intelligence we create is fundamental for the successful running of our organizations business.

Does this mean this new force will put limits on what we do? Or even worse, does it mean that we have to go over to the dark side to help our businesses? Absolutely not! The new law defines that we must gain permission, perform tests, and document. In no way does it restrict what honest data scientists can do. If anything, it provides us with opportunities.

JavaScript Nuggets on Demand

KNIME Analytics Platform is extremely flexible. It offers not only a number of pre-packaged functionalities for prototyping or routine work, but also a number of integrations for the free coding days. One of these integrations imports the power of JavaScript code into the platform.

This blog post series aims at providing nuggets of JavaScript code to implement more creative drawing and plots than what is already available with the pre-packaged nodes. The nuggets of JavaScript code proposed here implement only one functionality and are explained step by step for all, even the JavaScript beginners, to understand.

Today: Interactive Choropleth World Map using Google GeoChart visualization

Authors: Rosaria Silipo & Paolo Tamagnini

Figure 1. A choropleth map is a geographical map where areas are colored, shaded, or patterned according to a corresponding calculated measure, in this case logarithmic number of 2013 population on a world map.

The Plot

Today we want to draw the choropleth map as shown above. So what do we need?

• A map of the countries of the world and the corresponding numbers of population.
• A short JavaScript code to load the Google Charts library and draw the choropleth map based on the population numbers of each country.
• A Generic JavaScript View node to execute such code within a KNIME workflow.

Our dataset is the CSV file population2013.csv and it contains a list of 214 world countries with their corresponding population numbers as of 2013.

We also have a generic JavaScript View node. The smallest workflow would simply include a File Reader node to read the CSV file and a Generic JavaScript View node with the right JavaScript code nugget to draw the choropleth map. So let’s now have a look at this nugget of JavaScript code.

Hi! My name is Emil, I am a Teacher Bot, and I can understand what you are saying.

Remember the first post of this series? There I described the many parts that make me, or at least the KNIME workflow behind me (Fig. 1). A part of that workflow was dedicated to understanding. This is obviously a crucial step, because if I cannot understand your question I will likely be unable to answer it.

Understanding consists mainly of text processing operations: text cleaning, Part-Of-Speech (POS) tagging, tagging of special words, lemmatization, and finally keyword extraction; especially keyword extraction.

Figure 1. Emil, the Teacher Bot. Here is what you need to build one: a user interface for question and answer, text processing to parse the question, a machine learning model to find the right resources, and optionally a feedback mechanism.

Keywords are routinely used for many purposes, like retrieving documents during a web search or summarizing documents for indexing. Keywords are the smallest units that can summarize the content of a document and they are often used to pin down the most relevant information in a text.

Automatic keyword extraction methods are wide spread in Information Retrieval (IR) systems, Natural Language Processing (NLP) applications, Search Engine Optimization (SEO), and Text Mining. The idea is to reduce the representation word set of a text from the full list of words, i.e. that comes out of the Bag of Words  technique, to a handful of keywords. The advantage is clear. If keywords are chosen carefully, the text representation dimensionality is drastically reduced, while the information content is not.

Hi! My name is Emil and I am a Teacher Bot. I am here to point you to the right training materials for your early questions on how to use KNIME.

In a previous post, I described the KNIME workflow that built me. In that post the focus was on the details of my deployment (Fig. 1). A brain was mentioned, but very little explanation was given on the how-to for assembling it. The moment has arrived to give you some of these details and, in particular, to explain how my brain was conceptually designed to give answers.

Remember? It all starts with a question, your question. After parsing it and understanding it with the help of the KNIME Text Processing extension, queries are sent to my brain for possible answers.

Figure 1. Emil, the Teacher Bot. Here is what you need to build one: a user interface for question and answer, text processing to parse the question, a machine learning model to find the right resources, and optionally a feedback mechanism.

Educational Resources on the KNIME site

My ultimate goal is to provide you with the one and only web tutorial that perfectly answers your question. Let’s start then with the tutorial material that is available on the KNIME website.

First of all, of course, the e-learning course. The e-learning course consists of 7 chapters (so far). From installing KNIME Analytics Platform and its extensions to data access, from ETL procedures to Machine Learning, from data visualization to flow control: there is a good chance the topic you are looking for is here.

Another very much visited resource is the KNIME blog. The KNIME blog started in 2014 with the mission of producing nuggets of data science, machine learning, and KNIME tools.

Then, there is the Node Guide. Similar to the blog, this resource contains even more atomic pieces of information on single node and single use case usage.

Finally, there could be all the books from the KNIME Press, covering basic and advanced KNIME, text processing, data blending, ETL, etc …

Do you remember the Iron Chef battles? 

It was a televised series of cook-offs in which famous chefs rolled up their sleeves to compete in making the perfect dish. Based on a set theme, this involved using all their experience, creativity, and imagination to transform sometimes questionable ingredients into the ultimate meal.

Hey, isn’t that just like data transformation?Or data blending, or data manipulation, or ETL, or whatever new name is trending now? In this new blog series requested by popular vote, we will ask two data chefs to use all their knowledge and creativity to compete in extracting a given data set's most useful “flavors” via reductions, aggregations, measures, KPIs, and coordinate transformations. Delicious!

Want to find out how to prepare the ingredients for a delicious data dish by aggregating financial transactions, filtering out uninformative features or extracting the essence of the customer journey? Follow us here and send us your own ideas for the “Data Chef Battles” at datachef@knime.com.

Ingredient Theme: A Social Forum. Sentiment vs. Influence

Author: Rosaria Silipo & Kilian Thiel
Data Chefs: Haruto and Momoka

Ingredient Theme: A Social Forum

Today we have decided to go vintage and show the analysis implemented in the first KNIME whitepaper, where text processing met network analytics by Tobias Koetter, Kilian Thiel, and Phil Winters.

We propose the data from year 1999 of the Slashdot News Forum. Slashdot (sometimes abbreviated as “/.”) is a social news website, which was founded in 1997 for science and technology. Users can post news and stories about diverse topics and receive online comments from other users (cfr: Wikipedia).

Some years ago, we started a debate on whether the loudest customers were as important as everybody – including they themselves - thought. We started looking for public data on customer interactions about a given product and stumbled upon the Slashdot dataset. Users in the Slashdot data set are not strictly customers; they interact via a social forum about a given topic. If the topic were a product, they would be customers. So, assuming that talking about a product is a particular instance of talking about a generic topic, we decided to adopt the Slashdot data set for the analysis. We propose this same data set here again for today’s challenge.

Authors: Andisa Dewi and Kilian Thiel

Practically everyone has heard of Little Red Riding Hood, a fairy tale of an encounter between a young girl and the Big Bad Wolf. One of the most popular versions of this tale was written by the Brothers Grimm. But what has a simple folk tale got to do with text stream visualization? And what is stream visualization?

Stacked area charts, or stream visualizations, are very useful to show how topics in a single, or set of, documents change and develop over time. Each document is assigned to a single point in time, for example the publication date, and then specific topics or keywords can be visualized “stacked” on top of each other based on their frequency. A single chart is usually based on multiple documents.

So what do you think about reading the story of Little Red Riding Hood through a stream visualization? Taking the main characters of the story as our keywords, the visualization would show whenever the characters occur and co-occur, revealing their respective importance. This would give an idea of the course of the story without actually reading it!

The reason Little Red Riding Hood is good as an example is because it has a fairly simple storyline with only 5 main characters: Little Red Riding Hood, her mother, her grandmother, the Big Bad Wolf, and a Hunter.

Authors: Rosaria Silipo, Vincenzo Tursi, Kathrin Melcher, and Phil Winters

Hi! My name is Emil and I am a Teacher Bot.

I was built to answer your early questions on how to use KNIME. Pardon!

I was built to point you to the right training material to help you answer your early questions on how to use KNIME.

By the way, I was myself entirely built using KNIME. So, I should know where the right answers lie in the midst of all the tutorials, videos, blog posts, whitepapers, example workflows, and more, which are available out there.

It was not so hard to build me. You just needed:

• a user interface - possibly web or speech based - for you to ask your question
• a text parser for me to understand your question
• a brain to find the right training material to answer your question
• a user interface to provide the answer back
• a nice to have - but not necessary - feedback option, on whether my answer was of any help.

Regularization can be used to avoid overfitting. But what actually is regularization, what are the common techniques, and how do they differ?

Well, according to Ian Goodfellow [1]

“Regularization is any modification we make to a learning algorithm that is intended to reduce its generalization error but not its training error.”

In other words: regularization can be used to train models that generalize better on unseen data, by preventing the algorithm from overfitting the training dataset.

So how can we modify the logistic regression algorithm to reduce the generalization error?

Common approaches I found are Gauss, Laplace, L1 and L2. KNIME Analytics Platform supports Gauss and Laplace and indirectly L2 and L1.

Systems that automate the data science cycle have been gaining a lot of attention recently. Similar to smart home assistant systems however, automating data science for business users only works for well-defined tasks. We do not expect home assistants to have truly deep conversations about changing topics. In fact, the most successful systems restrict the types of possible interactions heavily and cannot deal with vaguely defined topics. Real data science problems are similarly vaguely defined: only an interactive exchange between the business analysts and the data analysts can guide the analysis in a new, useful direction, potentially sparking interesting new insights and further sharpening the analysis.

Therefore, as soon as we leave the realm of completely automatable data science sandboxes, the challenge lies in allowing data scientists to build interactive systems, interactively assisting the business analyst in her quest to find new insights in data and predict future outcomes. At KNIME we call this “Guided Analytics”. We explicitly do not aim to replace the driver (or totally automate the process) but instead offer assistance and carefully gather feedback whenever needed throughout the analysis process. To make this successful, the data scientist needs to be able to easily create powerful analytical applications that allow interaction with the business user whenever their expertise and feedback is needed.