Company data processing
Turning a public register into a working list
& filter Python processing
Starting with the question, then the dataset
I used Python to process a large public dataset of Estonian companies for website-development and SEO research. I started with two questions: which businesses might benefit from a website, and which already had a site that might need improvement?
The dataset needed narrowing before it could support that work. I grouped companies by business niche, excluded those identified as inactive and filtered on reported revenue or turnover in the previous year. That produced smaller sets that could be examined for a particular service, rather than one undifferentiated list.
Two routes from the same records
Companies without a website were possible development prospects. Those with a weak site were possible SEO prospects. The data preparation supported both searches, but the interpretation depended on the question being asked.
Reported turnover was a selection field, not a substitute for understanding the business. It did not establish profit or current demand. A company's place in a filtered list also did not mean it needed my services or wanted to buy them. Individual review still had to connect the record to a real need.
One dataset, two research paths
A useful handoff for further research
The outcome was a more manageable starting point for prospect research. The work demonstrates a defined data-processing assignment: organise records, apply explicit business criteria and keep the resulting groups aligned with their intended use.
I can prepare public data around an agency’s agreed research goal. It is useful before individual sales research or a website-review project, where cleaner segmentation helps people spend their attention on the relevant records.