Using Data Faster and Better
On better data visualization and modeling
This week, the DataSeries team held along the AWS team a Founders Dinner in Paris, in order to explore the startup ecosystem in the city. The attendees were not randomly selected, they all shared something in common — every single one was representing a data-intensive startup.
One of the concepts that resounded with me the most is that data should help make business decisions faster and more accurately. Many tools and models take too long, are hard to deploy and often offer limited solutions — isn’t it time to use data faster and better?
Data should be understandable
One of the founders, Shankar Sarul, was especially concerned in the way some traditional data analytics platforms like Tableau worked. He realized it while working and using this software at Groupon. A solution that was supposed to help businesses visualize how customers behave, was making life even more complicated — at some point, he had to get help or watch video tutorials to understand Tableau.
Sarul realized that there had to be another way to better access to customer insights in a visual fashion without spending days. That was the turning point that led Sarul to start Jetpack Data. The startup was born in 2016 with the vision to turn data visualization into an automatic solution.
JetPack claims to be the easiest and simplest way to view your data. By dropping the data into the platform and selecting the desired attributes, the user will get a working visual representation of the dataset. The main differentiator is the ease of use: it offers a walkthrough and you can get your results in a few minutes from the moment you get in.
Small data is meaningful data
Nowadays, a shared common thought is that Big Data is the best data, but I find that it depends on what the user needs to know. Small Data can also be very actionable data, and extract information that you might miss in the mess of Big Data. Big Data’s best use is for finding correlations, while Small Data’s best use is to find personalized insights.
If you don’t believe in the power of Small Data, ask Martin Lindstrom, who spends almost entire years going from house to house around the world studying tiny patterns. His verdict: Small Data has a huge impact on the stakeholder’s life. If a person that feels always great feels one day uneasy, looking for Small Data — a drop in regular exercise or sleep — might reveal quick insights that will help solve the problem.
MyDataModels — one of the attendees — has spent most of its existence building the engine that allows using Small Data to generate meaningful models. It can be applied to predicting house prices, preventing heart diseases or early detecting breast cancer. They can deliver more tailored solutions, but also faster — whereas Big Data models take weeks to months, their Small Data models take only days.
Tools that are either speeding up the modeling process, easing its use, or improving its accuracy, are truly a game-changer for users. The less that these tools take over users time, the more it will allow them to spend time on what truly matters — analyzing the insights and taking action.