Data science has become a critical tool for doing business in the uncertain economy. Here are 10 tools that can help you get and maintain that competitive edge.
Cynthia Harvey Freelance Journalist, InformationWeek on January 24, 2023
Most enterprise leaders recognize that data science and related disciplines are essential for competing in the modern economy. But many have struggled to mature and scale their data and analytics efforts.
According to IDC, organizations that are in the top quartile for enterprise intelligence (i.e., outstanding data science and business analytics capabilities) “are 2.7x more likely to have experienced strong revenue growth between 2020 and 2022, and 3.6x more likely to have accelerated time to market for new products, services, experiences, and other initiatives.”
Forrester refers to these organizations with excellent data science capabilities as “advanced insights-driven businesses.” And it noted that only 7% of companies met the criteria for that moniker in 2021. It predicts, “The decisions made in 2023 will fuel or extinguish a world of insights opportunity. With an uncertain 2023 approaching, data teams sit atop a tipping point that looks like rollercoaster carts gathering on a straightaway before the drop — only data teams with their partners, practices, and platforms lined up and prepared will move with speed and efficiency in the uncertain year ahead.”
Many teams hoping to achieve the necessary level of preparation are evaluating their current data science technology stack and considering making changes.
Today many teams are using a wide range of different tools. Gartner notes, “Analytics portfolios are becoming increasingly complex as a result of cloud migrations, new and disconnected ecosystems, and emerging self-service demands.” And it predicts, “By 2023, ease of migration, interoperability and coherence will be deciding factors in 90% of data science, machine learning, and AI platform buying decisions.”
So which tools will data leaders be evaluating when looking for interoperability and coherence and making those buying decisions?
This slideshow highlights 10 of the most popular data science tools available today. It includes data science platforms, programming languages, and other tools that can help enterprises become more data driven.
10. Trifacta/Alteryx
Trifacta is a popular data science tool that can speed up the process of data wrangling and preparation. Trifacta quickly converts raw data into a format that data scientists can use for actual analysis, process that would otherwise take a very long time. (Some say 80% of a data scientist’s time can be spent on these activities.) Trifacta works by combing through raw data sets, identifying potential alterations, and then automatically making transformations. By using Trifacta for data preparation and cleansing, data scientists are able to spend more time on actual data science-related problems. Trifacta was recently bought by Alteryx for 400 million dollars.
9. DataRobot
DataRobot uses artificial intelligence and machine learning to assist data users with data modeling. It aims to democratize the data modelling process, and it truly offers something for everyone. The platform is very easy to use and doesn’t require knowledge of programming or machine learning, so business analysists with little programming experience can build sophisticated predictive models. At the same time, it offers the tools for experienced data scientists and engineers to go deeper to produce even better predictive models. Datarobot is also very flexible and supports R, Python, H20, Spark ML, Vowpal Wabbit, and many others. DataRobot’s accessibility and flexibility along with its speed and reliability have helped to ensure its platform is widely used by both data scientists and non-data scientists all around the world.
8. SQL
While unstructured data stores get a lot of press, data scientists still do plenty of work with structured data that resides in traditional databases. And for accessing that data, they frequently rely on SQL (Structural Query Language).
In a 2020 Data Science Survey done by Kaggle, 44% of respondents said they regularly used some form of SQL. Many of them are querying data from SQL-based databases like MySQL, PostgresSQL, SQL Server, and SQLite, but you can also use SQL with big data tools like Spark and Hadoop. While it’s not a new or sexy technology, SQL provides easy, efficient access to structured data, and is an essential part of the data scientist’s toolbox.
7. Excel
Another of the most popular tools for data scientists is another of the lowliest and most overlooked — Microsoft Excel.
The ubiquitous spreadsheet application might not be the first tool that comes to mind when you think of data science, but it is one of the most widely used among data scientists for the purposes of data processing, data visualization, data cleaning, and performing calculations. Additionally, you can easily pair it with SQL to analyze data more efficiently. While not suitable for handling the enormous data sets that data scientists often work with, Excel is a great tool for performing data analysis on smaller scales and is a tool every data scientist should be familiar with.
6. SAS Viya
One of the most comprehensive data management and analysis platforms on the market, SAS Viya was created specifically for data analysis. It is one of the most popular statistical analysis tools among large companies and organizations, due to its great reliability, security, and ability to work with large data sets. SAS also offers extensive libraries and tools to assist data scientists in their data modeling and integrates with many popular tools and programming languages. It’s cloud-based and includes AI-based automation capabilities. However, due to its high cost, it is not widely used by smaller organizations
5. Tableau
One of the most widely used data visualization tools among data scientists, Salesforce’s Tableau can analyze large amounts of both structured and unstructured data. It can then take the data it analyzes and convert into a variety of helpful visualizations including interactive graphs, charts, and maps. What makes Tableau so useful is its ability to connect to a wide variety of different data sources. Tableau can easily connect to relational databases, file formats, and large cloud services like Azure and Google. Like DataRobot, Tableau is fairly easy to learn and use even by those without a programming background.
4. R
The R programming language is widely used for data science, more specifically for statistical modelling and analysis. Aside from Python, it’s probably the most important language to know for anyone working in data analytics. Data scientists use R and Python for very similar purposes, but there are a few key differences. To a greater degree than Python, R concentrates on the statistical aspects of data science. R performs more slowly, is more difficult to learn, and is less scalable than Python, but it is generally better when doing data visualization and analysis. It is open source and compiles and runs on most operating systems.
3. Apache Hadoop
Extremely popular for “big data” repositories, Apache Hadoop is an open-source framework for processing and storing enormous amounts of data. Hadoop works by distributing big data tasks across computing clusters. This is important because it allows an organization’s big data systems to operate in way that is both scalable and cost-effective. Additionally, it helps prevent widespread systems failures because if one node in the system goes down, Hadoop automatically redirects tasks to other nodes. Hadoop is standard among businesses that work with big data, so becoming familiar with it is crucial for anyone looking to get a job working with big data.
2. TensorFlow
Created by Google, TensorFlow is an open-source library for developing machine learning applications. Providing users with a vast array of resources and tools, TensorFlow is well known for enabling machine learning developers to build large and highly complex neural networks. Additionally, TensorFlow is highly compatible with Python, and its software libraries come full of many prewritten models to help with certain tasks. For example, TensorFlow can be used to recognize images, process natural language, and classify handwritten numbers and letters. Google Cloud and other cloud computing services offer services based on TensorFlow, which can make it easy to get started with the technology.
1. Python
Python has been by far the most popular programming tool among data scientists in the past few years. In the Kaggle survey 86.7% of data scientists said that they use Python, which was more than double the second most popular response. Python is relatively simple and easy to learn, which makes it easy for those without an extensive programing background to learn to read and write Python code. Many of the most popular data science tools are either written in Python or highly compatible with Python. Knowing Python is crucial for anyone working in data science as most data science jobs will require at least a basic Python background.