Data Wrangling

𝘏𝘪 𝘦𝘷𝘦𝘳𝘺𝘰𝘯𝘦 — 𝘚𝘵𝘢𝘳𝘵𝘪𝘯𝘨 𝘵𝘰𝘥𝘢𝘺, 𝘐'𝘭𝘭 𝘣𝘦 𝘳𝘦𝘨𝘶𝘭𝘢𝘳𝘭𝘺 𝘴𝘩𝘢𝘳𝘪𝘯𝘨 𝘴𝘰𝘮𝘦 𝘴𝘩𝘰𝘳𝘵, 𝘱𝘳𝘢𝘤𝘵𝘪𝘤𝘢𝘭 𝘪𝘯𝘴𝘪𝘨𝘩𝘵𝘴 𝘧𝘳𝘰𝘮 𝘮𝘺 𝘺𝘦𝘢𝘳𝘴 𝘢𝘴 𝘢𝘯 𝘪𝘯𝘥𝘦𝘱𝘦𝘯𝘥𝘦𝘯𝘵 𝘤𝘰𝘯𝘴𝘶𝘭𝘵𝘢𝘯𝘵, 𝘵𝘩𝘪𝘯𝘨𝘴 𝘐 𝘸𝘪𝘴𝘩 𝘮𝘰𝘳𝘦 𝘱𝘦𝘰𝘱𝘭𝘦 𝘵𝘢𝘭𝘬𝘦𝘥 𝘢𝘣𝘰𝘶𝘵. 𝘍𝘪𝘳𝘴𝘵 𝘶𝘱: 𝗱𝗮𝘁𝗮 𝘄𝗿𝗮𝗻𝗴𝗹𝗶𝗻𝗴.

We all love to talk about conclusions and insights from data, but nobody likes to talk about data wrangling.

Yet in my years of independent consulting, I've rarely walked into an engagement where the data was clean, centralized, and ready to go. More often, it's scattered across spreadsheets, systems, and siloed teams — inconsistently labeled, partially duplicated, and trusted by no one.

Before any analysis can happen, someone has to do the unglamorous work of finding it, cleaning it, and turning it into something reliable. That's data wrangling — and it's almost always the most underestimated part of any project.

Here’s why it matters more than people think:

1. 𝗠𝗲𝘁𝗿𝗶𝗰𝘀 𝗮𝗿𝗲 𝗲𝘃𝗲𝗿𝘆𝘄𝗵𝗲𝗿𝗲, 𝗯𝘂𝘁 𝘁𝗵𝗲 𝗼𝗿𝗶𝗴𝗶𝗻𝘀 𝗮𝗿𝗲 𝗻𝗼𝘁 𝗮𝗹𝘄𝗮𝘆𝘀 𝗮𝗽𝗽𝗮𝗿𝗲𝗻𝘁. Leaders know their Key Performance Indicators (KPIs) and the goals they are working towards, but oftentimes, there’s only a high-level understanding of how the measures are captured and calculated. Bringing it all to the surface can be genuinely illuminating. I once sat in a meeting where a data consolidation exercise revealed that two teams were measuring what they thought was the same thing completely differently.

2. 𝗜𝘁 𝗳𝗼𝗿𝗰𝗲𝘀 𝗮𝗹𝗶𝗴𝗻𝗺𝗲𝗻𝘁 𝗼𝗻 𝗮 𝘀𝗶𝗻𝗴𝗹𝗲 𝘀𝗼𝘂𝗿𝗰𝗲 𝗼𝗳 𝘁𝗿𝘂𝘁𝗵. Nothing exposes organizational misalignment faster than asking five people to pull the same number. Data wrangling doesn't just clean data — it creates the conditions for honest conversation.

3. 𝗖𝗹𝗲𝗮𝗻 𝗱𝗮𝘁𝗮 𝗽𝗿𝗼𝘁𝗲𝗰𝘁𝘀 𝘆𝗼𝘂𝗿 𝗰𝗼𝗻𝗰𝗹𝘂𝘀𝗶𝗼𝗻𝘀. If your findings are inconvenient for someone in the room, they will look for any error they can find to discredit the whole analysis. Clean, well-documented data removes that escape hatch.

𝗢𝗻𝗲 𝗯𝗲𝘀𝘁 𝗽𝗿𝗮𝗰𝘁𝗶𝗰𝗲 𝗜 𝗳𝗼𝗹𝗹𝗼𝘄: 𝗯𝗲𝗳𝗼𝗿𝗲 𝗰𝗿𝗲𝗮𝘁𝗶𝗻𝗴 𝗮 𝘀𝗶𝗻𝗴𝗹𝗲 𝗳𝗼𝗿𝗺𝘂𝗹𝗮, 𝗺𝗮𝗽 𝗼𝘂𝘁 𝗲𝘅𝗮𝗰𝘁𝗹𝘆 𝘄𝗵𝗮𝘁 𝘆𝗼𝘂 𝘄𝗮𝗻𝘁 𝘁𝗵𝗲 𝗳𝗶𝗻𝗮𝗹 𝗱𝗮𝘁𝗮𝘀𝗲𝘁 𝘁𝗼 𝗹𝗼𝗼𝗸 𝗹𝗶𝗸𝗲. Every column (including source data you want, as well as derived information that will be calculated from source fields). Every row. Then figure out how to fill it in the cells (and make sure you have an unique ID for each row; more on this in the future).

The analysis is the exciting part. But it only works if the foundation is solid.

#consulting #datastrategy #businessintelligence