From understanding client’s needs, to learning how to clean and reshape data, my first week at The Data School has been hugely rewarding.
Starting with client needs, my biggest takeaway is that context is key. It is pivotal to understand what your client does, what is happening in their industry, and what questions they are trying to answer. During any kind of placement or project, continual engagement with your stakeholders allows you to build rapport and gain a true understanding into what they are trying to achieve. Researching the investment platform Freetrade was a great insight into how due diligence could be conducted.
Example: Women’s Figure Skating Dashboard
Figure Skating is an event at the Youth Olympics, where competitors conduct a routine and are ranked by judges. Each competitor is ranked during qualifiers, then awarded a ranking for the finals.
The dashboard shows how each routine differed in terms of the type of elements performed at each stage, and clearly shows that everyone’s Finals routine was notably longer than their Qualifier routine. My job was to ask questions of this dashboard, thinking of what could be asked by various stakeholders. For instance, the perceived ranking/difficulty of these elements would be clear to someone familiar with Figure Skating, but to a “layman” who might be investigating what a successful routine looks like, some supplementary information regarding these elements could be helpful.
Data Preparation
Moving onto what we learnt about Data Preparation, learning the differences between Databases, Data warehouses and Data Lakes answered all my questions on how and where data is stored and sourced from. Learning cleaning and reshaping techniques for Tableau Prep such as pivoting rows to columns and splitting columns was great preparation for the first Friday Project. I was tasked with profiling a dataset on complaints to a company providing Financial Products. I then presented what my plan was to reach a desired output, showing in detail how I profiled the dataset and what questions I had of the data.
Above shows my plan. I summarised the main dataset I had, what supplementary tables were provided and what steps I’ll be taking to reach my desired output, along with a sample of this. The yellow boxes show what steps need to be taken in Tableau Prep to reach this stage.
Below is the data profile of the initial dataset, showing what types of data I had, some interesting “quirks” of the dataset, and some questions that need to be answered by the stakeholder.
Key Takeaways & Lessons Learnt
- Time management! Absolutely key to prioritise your time. It is much better to have a clear plan of your workflow, than have output that does not show a thorough understanding of the data or the client’s expectations. I spent far too much time thinking about how I’d reach my desired output, trialing things on Tableau Prep instead of taking ample time to properly scrutinise the dataset and devise a clear plan.
- Always Be Communicating. Had I properly assessed the dataset from the beginning as opposed to later in the day, I would’ve had a lot more time to go back to the stakeholder and have my questions answered. By not having an ongoing engagement with my stakeholder, I left a lot to be answered which could easily have been avoided.
- Document Everything. Post-presentation, I was asked a lot of questions, some of which I did not document the answers to, and could have avoided confusion by documenting these things. In this case, I did not initially include information about data from different years being structured differently in terms of fields.
Final Thoughts
Overall, I thoroughly enjoyed my first week. I look forward to applying these fundamental skills I have learnt to future projects, and delivering my first presentation in a long while has shown me how much room for improvement there is in this regard.