instapipe

The problem

Meta Business Suite exports in its own format with its own date logic, column names and metric definitions — which do not match TikTok's even when measuring the same thing. The real problem is not parsing the CSV: it is that "reach" on Instagram and "unique views" on TikTok are the same concept with different names, and if you do not normalise it in the pipeline, every analysis normalises it differently.

The data

Meta Business Suite exports: reach, impressions, interactions by type (likes, comments, saves, shares), daily follower evolution and stories performance. The date format uses a localised month in the column name (different between the Spanish and English app versions).

Approach

• Parser that handles the Meta Business Suite format including the localised month in column names
• Semantic normalisation: explicit mapping of each Instagram metric to its tokpipe equivalent
• Output with the same columns and types as tokpipe so the two platforms' DataFrames can be concatenated directly

Tools

python

pandas

Result

Direct complement to tokpipe. Together they allow analysing TikTok and Instagram with the same metric definitions, without each analysis resolving the normalisation ad hoc.