What Data Science Tools Have you used in the past 12 months? KDnuggets Poll Results are out!

The results of the 18th annual KDnuggets Software Poll were recently published. This poll asks “What Predictive Analytics, Data Mining, Data Science Software/Tools have you used in the past 12 months?”.  This poll attracted 2,900 voters, and it is also worth mentioning that it sometimes attracts controversy due to excessive voting by some vendors.  See all data at KDNuggets.

Some of the most relevant findings are:

  • Python has now overtaken R as a Data Science Tool – barely but still noticeable (53% to 52% use but Python grew 15% while R only grew 6%)
  • There are now 2 newcomers joining the top 10 list: Tensorflow and Anaconda
  • Use of Excel for Analytics purposes decreased by 16%
  • In terms of programming languages, Python, R and SQL run the show with usage of all 3 growing
  • Big Data Tools was simplified to only 4 categories: Hadoop Open Source, Hadoop Commercial, SQL on Hadoop Tools and Spark.  The highest growth tool is SQL on Hadoop and usage of Hadoop Open Source is decreasing


We have 2 newcomers this year: Anaconda and Tensorflow

Top 2 tools:

  • Use: Python (53%) and R (52%)
  • Growth: Tensorflow (197%), Anaconda (37%)


Top 2 languages:

  • Use: Python (53%) and R (52%)
  • Growth: Python (15%), R (6%)



The tools on the survey have been simplified to 4: Hadoop Open Source, Hadoop Commercial, Spark and SQL on Hadoop Tools.

Top 2 Big data tools:

  • Use: Spark (23%) and Hadoop Open Source (15%)
  • Growth: SQL on Hadoop Tools (41%), Spark (5%)

It is important to notice the decrease of 32% in usage of Hadoop Open Source. I am not sure if there has been a real decrease, or if this is the result of the survey having changed splitting the hadoop category in 2: Open Source and Commercial.  Part of this “decrease” could be attributed to the fact that there are now 2 categories instead of 1.


Top Deep Learning Tools:

  • Use: Tensorflow (20%) and Keras (9.5%)
  • Growth: Microsoft CNTK (278%), mxnet (200%)

What is Cognitive Computing? The 3 Things Series

*** The 3 Things Series aims to simplify – sometimes even oversimplify – technology concepts so that you learn 3 things about a topic ***. Opinions are my own.

Cognitive Computing is an evolving technology that –  although in its infancy – has the potential to revolutionize all industries, by redefining how work gets done, and by augmenting human capabilities.

In a general sense, cognitive systems – with IBM Watson at the forefront of this new era – combine natural language processing, machine learning, and real-time computing power to process vast amounts of data – structured and unstructured – to provide answers to a specific request.

These systems are very different from traditional systems that are programmed based on rules and logic. They enable people to create value finding insights in volumes of data, while mimicking cognitive elements of human expertise. They “learn” about a specific domain, develop hypothesis, evaluate those hypothesis, and choose the best option. They do this at massive speed and scale.

So what does this all mean? I find that the best way is to show you the origins of IBM Watson. If you watch this 3-minute video, you will have a better understanding: http://bit.ly/2qhElVe

So how does a cognitive system work?

1.    The system first needs to be trained with domain specific data. If we think of jeopardy, Watson had access to about 200 million pages of structured and unstructured content including the full text of Wikipedia. This stage is very important as the outcomes will be as good as the information used in the training process. In fact, when Watson was being developed, the urban dictionary was included as part of the training only to be removed later when pretty quickly it learned how to use offensive words. If we think about the field of oncology for example, then the training data will include information on research papers, drugs used, demographics data on individuals receiving training, outcomes, etc.  So it must be able to understand both structured and unstructured data, but also to interact with humans in a natural way (through natural language)

2.    Once a system is trained, it will have the ability to form hypothesis, make considered arguments and prioritize recommendations to help humans make better decisions.    Cognitive systems are probabilistic, and they generate responses according to levels of confidence. They can also show the evidence for the responses—what data there is to back up the answer and the confidence score. If we think about the jeopardy scenario, you probably noticed on the video above that the top 3 options together with their confidence score were displayed on the screen. Decisions can then be made based on those confidence scores to select the best option. In the oncology example, the ability by a doctor to look at all the evidence the system used, and at the collection of hypothesis that could potentially include some that the doctor had not considered before, is extremely valuable.

3.    Cognitive systems ingest and accumulate data insights from every interaction. The confidence levels it provides are subject to change when subject matter experts grade the responses since the system is not programmed but trained by experts who enhance, scale and accelerate their expertise.

So what are the 3 things to remember about cognitive systems? They are trained, not programmed, they provide probabilistic responses with confidence levels instead of exact answers, and can get better overtime as experts enhance, scale and accelerate their expertise.

What is Analytics – The 3 Things Series


Before I embark in explaining what is Analytics, as well as the different types of Analytics, let’s just talk for a second about Why Analytics. The field of Analytics was born with the goal of using data and the analysis of that data to improve performance in key business domains, which basically means having the ability to make better decisions, and to execute the right actions based on data insight.

So what is Analytics?

The field of Analytics involves all that is necessary to drive better decision making and add value, such as Data platforms (On Premise, On Private or Public Cloud), Access to Data (Structured and Unstructured), and Tools for Quantitative Analysis and Data Visualization. In other words, Analytics is all about turning data into insight which in the world of business means turning data into competitive advantage.

Analytics Types

There are three main types of Analytics:

1- Descriptive Analytics help you understand “What happened?”. The goal of descriptive analytics like the name implies is to describe or summarize raw data and turn it into something that makes sense to the human eye – typically through presenting the data in tables or reports, or in visualizations or charts. They are very useful for understanding past behaviors, and how the past might potentially influence future outcomes.  Basic statistics like averages, sums, percent change, or proportions fit into descriptive analytics. This is the simplest form of analytics but nevertheless it is extremely useful and necessary as a stepping stone into more sophisticated/valuable analytics.

2- Predictive Analytics help you understand “What could happen?”. There are two main goals: Finding relationships or patterns and predicting what could potentially happen. They help you try to understand the future while providing actionable insights based on data. Predictive Analytics don’t provide predictions that are 100% accurate, but provide estimates about the likelihood of a future outcome. They can be used throughout an organization to forecast sales or inventories, to detect fraud, to understand customer behavior, or in any scenario where relationships among data and “predicting the future” will help make better decisions. We are all familiar with our credit scores right? That is an example of predictive analytics where historical data on how well you manage your credit is used to predict a score that can then be used as a proxy for how much of a credit risk you might be.

3- Prescriptive Analytics help you determine “What should we do?”. Prescriptive Analytics are all about providing specific guidance about what to do. They make an effort to quantify the effect of future decisions looking at the possible outcomes of each scenario before the actual decisions are made. They use a combination of business rules, algorithms, and modelling procedures to provide possible outcomes. These are typically used in Supply Chain Management, Price Optimization, Workforce planning among others. They are very useful like the name implies to “prescribe” a direction after examining multiple possible scenarios.

In summary, Analytics help you turn your data into insight for better decision making and there are 3 main types of Analytics that you use depending on your goal. Descriptive analytics to understand what has happened in the past, Predictive analytics to understand relationships among data and provide predictions about what may happen in the future, and Prescriptive analytics to provide specific recommendations about what to do in specific scenarios.

United’s CEO Fails to Understand the Power Shift brought on by Social Media


By now everyone has seen the despicable way a United Airlines passenger was treated, forcibly being removed from a plane to release his seat to another passenger. I personally find the whole situation appalling but that is not what I want to discuss now. I want to talk about the way United’s CEO handled the situation, and how it clearly demonstrates a failure to understand how social media has shifted power from the few to the many.

Imagine if this situation had happened 20 years ago, before a world where social media was ingrained in our lives. Even if people on the plane had been able to take pictures or record video, what choices did they have to share them? The result would have been that only a handful of people would have heard the story. The CEO and executives in turn, would have also had a lot of power to control the message, and probably would have been able to get away with this without losing several million dollars in market valuation due to people’s outrage around the world.

But in fact what happened was that within minutes of this event, people around the world were seeing video and pictures of this physical assault, and people were waiting for, or in fact, expecting a statement from United’s leadership. So after the first failure at managing the overbooking situation, came the second: the CEO’s response.

  1. He apologizes for “re-accommodating” a passenger, assuming that would be the end of the history
  2. He proceeds to blame the passenger accusing him of being belligerent and indicating it is important to find out why “the passenger acted the way he did”. He also doubles down on congratulating his employees for a job well done. Tone deaf much?
  3. After losing almost a $1B in valuation at some point during the day he finally comes out with the statement he should have issued from the beginning stating “He is sorry. This shouldn’t have happened, and they will take measures for this not to happen again”.

Lesson to learn from this event? Corporations – and their leaders – cannot get away with a lot of the things they could have gotten away with in the past, and we have social media platforms to thank for that. Ideally, business leaders would care about their clients and their business, but even if as a leader truly in your heart you don’t care and your first reaction is to say “Not our fault. We did everything by the book”, know that most likely that statement is only going to amplify existing outrage.

Yes, you may think it is possible to do it seeing how some politicians get away with so much gaslighting these days, but they have something you don’t, and that is followers willing to be gaslighted because they are blinded by their passion for a political party. More likely than not your clients and others don’t have that level of passion for your business.

So next time something like this happens…. Stay away from the temptation to blame the victim and address the situation the right way, which includes any combination of:

  • We are sorry
  • The buck stops here
  • We are investigating
  • We are taking steps to ensure this doesn’t happen again

Don’t forget: social media has shifted the power from the few to the many.

What is the Business Value of Big Data? – The Three Things Series

*** The 3 Things Series aims to simplify – sometimes even oversimplify – technology concepts so that you learn 3 things about a topic ***. Opinions are my own.

Organizations embark in Big Data projects typically with 3 goals in mind: cost reductions, improved decision making and the ability to create new products and services.

big data business value

1- Cost Reduction 

As the quantities and complexity of data in organizations increase, so does the cost of storing and processing this data. Decisions about how much data to keep available for analysis, and how much “historic” data to move to tape or other less expensive resources, are then made. The problem with this strategy is that by limiting the data that can be analyzed, the insight that can be derived from this data is also limited.

In recent years, technology developments especially in Open Source, have made cost reduction a reality through the use of inexpensive technology such as Hadoop clusters (Hadoop is a unified storage and processing environment that allows for data and data processing to be distributed across multiple computers). Hadoop clusters give organizations the ability to keep more data available for analysis at a lower cost, and to easily add complex data types (images, sound, etc) to the pool of data to be analyzed

2- Improved Decision Making

Data analysis can be significantly improved by adding new data sources and new data types to traditional data. For example a data-driven retailer may see significant benefits in their inventory planning processes, if a new data source like weather data is added to the model to better predict sales and inventory requirements. An enriched model may be able to predict shortages of winter clothing by incorporating temperature into the existing models. Additional benefits can also be achieved, if more complex data is analyzed. For example, this same retailer may better target their ads in social media, if they evaluate not only their clients purchasing history, but also the actions they take in social media to interact with their brands and those of their competitors.

3-  Development of New Products and Services

The most strategic and innovative business benefits will probably be achieved by the ability to use new data or new sources of data to create new products and services. Let’s think for a minute about the data our cars generate (yes, we don’t necessarily see it, but more and more cars are equipped with sensors that collect a lot of data about our driving history). Using this data, insurance companies can offer policies that are dynamically priced based on an individual’s driving history (which is good news to you only if you are a safe driver of course).   Integrating weather data can also bring tremendous savings to an insurance company. Some insurance companies have been able to achieve significant savings per claim by letting their clients know that a storm is coming and recommending they don’t leave their cars exposed to the elements. (Again, assuming that as a client you listen to your insurance company recommendations).

In summary, when thinking of the business value of Big Data, think of  three areas of value:

  • Cost reductions
  • Improved decision making
  • Ability to create new products and services


What is Big Data? (The 3 Things Series)

*** The 3 Things Series aims to simplify – sometimes even oversimplify – technology concepts so that you learn 3 things about a topic ***. Opinions are my own.

The technology industry is full of “buzzwords”, with Big Data being one of the most used in recent years. Organizations have always dealt with data and have stored that data in databases, but we can see in the chart below how searches on Google have changed throughout the years comparing searches for “Databases” to searches for “Big Data”.


databases vs big data searches


Big Data in general refers to the ability to gather, store, manage, manipulate, and – the most important one – get insights out of vast amounts of data. And the typical question is “how big does data need to be so it is considered Big?” And the answer is…. it depends. When it comes to size, an organization’s Big Data may be another organization’s small data.

There are 3 things to remember that define “Big Data”:

  • Volume. It refers to size. So if you are capturing vast amounts of information, you probably have Big Data in your hands
  • Velocity. Are you working with data at rest? Or data in motion? For example if you are analyzing sales figures for the past year, that data is at rest (it is not changing constantly). But if on the other hand you are analyzing tweets to understand how your clients are reacting to a product announcement, this is data in motion as it is continuously changing. It may not be necessarily big if you are looking at daily data, but the fact that it is data in motion is relevant to the definition of Big Data
  • Variety. As the ability to capture, store and analyze more data has increased, so has the interest in analyzing data that is more complex in nature. For example, an insurance company may want to analyze the recordings of customer service calls to determine what characteristics of the conversation led to a policy sale, a retailer may want to analyze videos to determine how people navigate the store and how that impacts sales, or a hospital may want to analyze x-rays to find patterns and correlations between common symptoms in patients.

So when it comes to the definition of Big Data, remember 3 things, or the 3 Vs:

  •  Volume (size)
  • Velocity (Frequency of data update during analysis)
  • Variety (complexity of data to analyze – images, videos, texts, log files, etc)




Enterprise Business Intelligence Platforms – At a Glance

Industry analysis reports are always packed with great information, lots of it.  Data visualization however helps better understand the conclusions reached by Forrester in their The Forrester Wave™: Enterprise Business Intelligence Platforms, Q1 2015 report.  11 vendors in this industry are included.

Table 1 shows the summary scorecard, and details are presented below the table.

Table 1 - Enterprise Business Intelligence Scorecard

Table 1 – Enterprise Business Intelligence Scorecard

The top level comparisons involve 3 categories: Current Offering, Strategy, and Market Presence.

The Forrester Wave™: Enterprise Business Intelligence Platforms, Q1 2015

Vendors with above-average scores

Current Offering: IBM, Information Builders, Microsoft, Opentext, SAP, SAS

Strategy: IBM, Microsoft, Oracle, SAP, and SAS

Market Presence: IBM, Microsoft, Oracle, SAP and SAS

We can also look at the scores that went into every one of these categories and see how vendors compare at a more granular level.

Current Offering Details

Forrester Evaluation Business Intelligence Current Offerings

Vendors with above-average scores for Current Offering

Architecture: Information Builders, Microstrategy, Opentext, SAP, SAS

Development Environment: Microsoft, Opentext, and SAS

Functional Capabilities: IBM, Information Builders, Microsoft, SAP and SAS

Operational Capabilities: IBM, Information Builders, Microstrategy, Opentext

Strategy Details

Forrester Enterprise Business Intelligence Strategy Comparisons

Vendors with above-average scores for Strategy

Commitment: IBM, Microsoft, Oracle, SAP, SAS

Pricing: IBM, Information Builders, Microsoft, Opentext, Qlik and TIBCO

Transparency: Tableau

Product Direction: IBM, Information Builders, Microsoft, Oracle, SAP, and SAS

Market Presence Details

Forrester Enterprise Business Intelligence - market presence evaluation

Vendors with above-average scores for Market Presence

Company Financials: IBM, Microsoft, Oracle, Qlik, SAP, SAS

Global Presence Base: IBM, Microsoft, Microstrategy, Oracle, SAP, and SAS

Partnership Ecosystem: IBM, Microsfoft, Oracle, SAP, and SAS

Functional Applications: IBM, Oracle, SAP, and SAS

What Data Science Tool Have you used in the past 12 months?

The results of the 16th annual KDnuggets Software Poll were recently published. This poll asks “What Predictive Analytics, Data Mining, Data Science Software/Tools have you used in the past 12 months?”.  This poll attracted 2,800 voters, and it is also worth mentioning that it sometimes attracts controversy due to excessive voting by some vendors.

93 different tools were included in the poll. To determine which tools are included, they start with the companies in Gartner Magic Quadrant(tm) for Advanced Analytics and Forrester Wave(tm) for Big Data Predictive Analytics, and add companies/tools from last year poll, and relevant new ones in the market.

R is the tool most frequently used tool in this community of data scientists, but other tools are growing rapidly (Spark, KNIME, Python).

Top 10 data mining tools


In terms of programming languages, Python is the clear leader.



And Hadoop is the clear leader in the Big Data Tools space.

big data tools


Women in Software Engineering – Data Visualizations

Tracy Chou, Software Engineer at Pinterest has become a leading voice for women in the tech industry by using data to call attention to how few of them are employed as engineers. She has uploaded a spreadsheet (https://github.com/triketora/women-in-software-eng), that companies can use to make public the number of female engineers in their ranks. The goal: to identify the scope of the problem as a first step toward making a stronger commitment to address it.

I used the data to create some visualizations.

Quantifying Silicon Valley’s Diversity Issue