Applying the 80/20 principle to the Data Analyst’s workday
What are the few key skills that you will need to cover 80% of your day-to-day activities as a data analyst?
With the boom of Data, the business has realized the importance of leveraging it for better business outcomes, and data related roles including that of a Data Analyst
With this came the buzzwords and different specialized tools for smaller parts of the Analytics pipeline. It becomes increasingly difficult to cover these, and you should not. I firmly believe that if we have a strong foundation, we can easily upskill these tools and deliver results.
Another thing that makes it complicated is the definition of data analyst in different companies. You might have the designation of a data analyst but that will encompass roles from business analyst, QA, data engineer, data scientist, Business Intelligence Engineer, or full-blown software engineer (trust me I have seen this).
Let me take the easy way out and say, what you need to do daily as a data analyst will depend on your organization.
But I will share my experience from various clients/businesses using different tech stacks to define what is expected from a data analyst. This is one of the perks of working as a Consultant, you get to experience lots of projects with varying tech stacks.
Case 1: Working for Food and beverages company as a Data Analyst
Tech Stack | Qlik Sense, Excel
This project was all about the migration of financial reports built in excel and legacy financial reporting systems to Qlik
Most of the time here went by in analyzing the current reports, defining business glossary(you guessed it, we didn’t have documentation for business metrics used by the client), writing Qlik expressions based on calculations in excel, testing, and validating reports to finally pushing it to production.
Case 2: Working for a TV DTH service provider
Tech Stack | Microsoft SQL Server, Tableau
The use case here was to create Tableau reports for business which was currently using excel for customer movement between different plans (upgrade, degrade). There was an interesting use case related to customer churn as well.
Apart from the advantage of having interactive operations in a tableau report compared to excel, we also added a few KPIs that were adopted by the business team
Case 3:Working for a B2C company proving laundry services
Tech Stack | Azure synapse Analytics, Tableau, Qlik Sense
The project here was migrating a few reports from Tableau to Qlik while creating new reports in Qlik for marketing and revenue teams. (no excel this time).
Although we had a Database schema (name DWH), it wasn’t, we had to create a reporting layer on top of it, optimized for reporting. And then taking Ad-Hoc requests and creating reports. The most interesting part here was experimenting with advanced analytics features in Qlik (like Forecasting and Clustering) for a few reports.
Although it might sound a bit daunting (if you are new to this), all these implementations were a few google searches away.
Conclusion
80% of the functions, expressions and SQL queries used in all three projects are what you learn initially, in the first 20%. And along the way when I got a few roadblocks, I used community for the corresponding tools to ask questions and was grateful as I almost always got a solution.
As I worked through different tools (Tableau, Qlik, and in a few projects Power BI as well) I barely got time to gain expertise over the tools. But as these tools are similar in many aspects, the knowledge gained in one BI tool is reusable across all, with few tweaks and modifications.
As I gained experience working on different projects, I took better design decisions, looked for performance improvement techniques, and tried to increase user adoption of the reports.
And we continue our learning, this 20% starts adding up.
A sample peak of enlightenment.
As each project is different, you might have to learn and upskill based on the project requirements (mashups in Qlik, embedding in Tableau, custom viz are challenging to learn, but equally rewarding). That said, 20% will get you by!
