So as I have explained in the first post in this "data to the people" series, I really think that anyone who wants to do data analytics and isn’t afraid of numbers actually should do data analytics. Related to the wisdom of the crowd paradigm, my opinion is that more people doing data analytics will be better than less.
In this post I'd like to give the first few concrete examples of how to achieve this, specifically within a company.
At Jimdo we started by giving analytics classes to colleagues who were interested. We as a datateam provide a self service data warehouse based on aws redshift, which is basically an extract from our base data warehouse. Instead of continuously working on a series of adhoc-queries (e.g. to find out how many paying user we have in a specific city) we decided to give our colleagues the tools to answer these queries autonomously themselves. We didn't want to be the data princelings who act as gatekeepers to the valuable data-treasure that we have, as Avinash Kaushik has described in this highly recommended talk.
Most importantly these analytics classes we really hands on, i.e. we didn't try to give a tutorial on SQL or even go into the details of relational databases. Instead we tried to be as practical a possible. We only worked with examples of queries that other colleagues have actually worked on.
Before we started with this Self Service SQL and data analytics classes we didn't know at all what the adoption-rate would be and how many of our non-technical colleagues would actually start working with SQL regularly. The results were quite intriguing: as expected, some colleagues who participated in our data analytics classes never afterwards worked with the datawarehouse. But many did and started integrating database queries in their daily workflow. There's even a repository of most-used sql-queries which is being managed by a group of non-technical country managers. One of our co-founders who has never before worked with SQL and is more of a business-type just published a report on how our mobile app usage looks like onto our company's yammer. Guess where he got the data from?
Additionally to the Beginner's SQL classes we introduced an "SQL Lean Coffee". Two members of the data team would be available for about an hour to go through specific SQL questions which are then being discussed with all attending colleagues. This works quite well, allows for open discussions and complements the standard classes.
One of the motivations for starting with Self Service SQL and these analytics classes was to be able to deal with the immense amount of adhoc queries that flooded the data team. By now enabling many colleagues to directly query our data warehouse is one of the cornerstones to make Jimdo truly data driven.
In the next post I will explain how we are developing more advanced analytics classes for our colleagues with existent SQL skills. Quite a few of our software developers have shown interest in topics such as data visualization and cluster analysis.
In the mean time, If you want to know more details about these analytics classes, think about doing them in your organization or just want to exchange your thoughts and experiences please feel free to contact me.