I swear if I see one more portfolio project analyzing Titanic survival rates, I\u2019m going to start rooting for the iceberg. <\/p>\n
In actual work, 80% of the job is cleaning messy, inconsistent, incomplete data. But every public dataset I find seems to be already scrubbed within an inch of its life. Missing values? Weird formats? Duplicate entries? <\/p>\n
I want datasets that force me to: My problem is, most companies won\u2019t share their raw internal data for obvious reasons, scraping can get into legal gray areas, and public APIs are often rate-limited or return squeaky clean data. <\/p>\n The difficulty of finding data sources is comparable to that of interpreting the data. I\u2019ve been using beyz to practice explaining my data cleaning and decision, but it\u2019s not as compelling without a genuinely messy dataset to showcase. <\/p>\n So where are you all finding realistic, sector-specific, gloriously imperfect datasets? Bonus points if they reflect actual business problems and can be tackled in under a few weeks.<\/p>\n<\/div>\n submitted by \/u\/Various_Candidate325 <\/a>
– Untangle inconsistent date formats
– Deal with text fields full of typos
– Handle missing data in a way that actually matters for the outcome
– Merge disparate sources that almost<\/em> match but not quite <\/p>\n
[link]<\/a><\/span> [comments]<\/a><\/span><\/p>
