IPL: Ways to see beyond Scoreboard - Part 2

Finally, it’s completed. The visualization is live now - visuals → ipl2025.

The learning from this activity is quite a long list.

1. Visualization is easier compared to data collection & preparation.
2. UI is scarier than the backend stuff like terminal or notebooks.
3. Prototyping is very much important, especially for the hobby project.
4. Iterative improvement is greater than aiming for a polished shiny thing.
5. Presentation is key

1. Visualization is easier compared to data collection & preparation.

If you look at the chart it’s a combination of

• circles,
• rectangles,
• some texts, and
• tooltip

That’s not a lot. That’s not at all complex to create. For the powerful tool like D3, the chart doesn’t even do justice. It is all about stacking the well-known shapes one over another to get some visuals.

What really bothered is how to obtain and prepare the data. There are not many APIs available to fetch Cricket results, that too for free. After chasing quite a number of APIs finally settled with cricSheet. The site has detailed info about every cricket match.

Second part is to prepare the data as per our requirement. Here’s where the Python comes in. Initially I used the Google collab to design a barely working code. Then slowly cleaned it to reuse the code every week, that’s the frequency with which the cricSheet gets refreshed. Pandas is what I usually employ at work. Polars was getting popular and so I used it in place of Pandas. The syntax and style is more similar to Pyspark, turns out much easier to get it done.

My day job helped me out with Data collection & preparation related activities. If not for the Data Engineering things that I learn in day job this chart couldn’t be possible.

2. UI is scarier than the backend stuff like terminal or notebooks.

My go to tools involves SQL, Python, some IDE, and terminal. Databricks notebooks are the closes thing to be a working UI. So, you can understand where in the UI spectrum I lives in. Data Visualization is diagonally opposite to my day job, and that’s what attracts me towards.

Learning or reading the stuff is much, much easier than doing. Easier said than done makes sense. But UI I feel should be learnt through doign than reading, or seeing some Youtube videos. There are so many changes made on the UI -

• Color combination - Early 2000s websites were much better than my very first colour combo
• Order of the matches - Until the semifinals the matches were in ascending order, the finals must be at the last of the chart, if I continued with same. But scrolling 2 pages to view the result doesn’t looks good, especially during this TikTok inspired attention period
• Tooltip - Earlier tooltip was not at all setup. So, verifying whether I’m projecting right details or not became an activity on its own.

3. Prototyping is very much important, especially for the hobby project.

If you’re doing a project to a firm, there must be someone else who had done a PoC and you just follow it. There could be someone in the hierarchy to help you out. But this is your hobby project. You’re the architect, developer, manager and the very first customer. So, to convince all these characters you should have a prototype and give a confirmation that whatever living in your head can be bought to the world. That’s the only way I see motivating myself. With money out of table, self satisfaction is the only motivation.

Initial time is when the motivation is very low. Setting up the Python environment and installing required libraries may not have helped. But, Google Colab helped me then. IT had working compute, easier to install libraries and get things done.

4. Iterative improvement is greater than aiming for a polished shiny thing.

I procrastinate a lot. Like a lot. Half my awake hours is spent living in my head. So many ideas that forms gets withered away while I build it in my head, looking for the finished project. This chart is different but same.

• Different - This chart is completed.
• Same - The imaginative version of chart started few years back.

This time its gets completed because I started doing instead of imagining various things the chart can potentially have. The initial versions were disturbing to see. Then slowly the chart reached this version of relatively better visuals.

5. Presentation is key

Whatever we make, the end product must be presented in a near manner for other to view it, and spend some time on it.

The chart is somewhat okayish. But the page doesn’t accommodate the entire chart, it needs to be scrolled right to have a good view. It’s a showstopper even for me. Also, the commenting section is not aligned with the chart another one to affect your flow.

This made me to think should I really build the page for data visualization myself instead of using the Hugo themes directly. Some thing to ponder over.

The Repo

The git repo that’s holding the underlying code is dheepakg/ipl-2025.

Credits

• cricSheet - For Cricket match related details.
• Google Colab - For prototyping data preparatory work

See also

• IPL: Ways to see beyond Scoreboard
• Visualization: Post grouped by years - D3 - v3
• Visualization: Post grouped by years - D3 - v2
• Visualization: Post grouped by years - D3
• Visualization: Post grouped by years