Browse categories
Explore
Fiverr Pro
English
$
USD
I will audit your python data analysis and deliver reproducible tests
About this Gig
Make your Python analysis easier to trust and rerun. I hold a PhD in Computational and Applied Mathematics and work in scientific computing and machine learning. I review an existing CSV-based analysis or Python notebook, check agreed data rules and calculations, and deliver a rerunnable test suite with an evidence-linked findings report. Packages cover bounded CPU-based Python workflows; send a representative file and your expected outputs before ordering. Deliverables include editable code, test results, environment notes and a concise report that separates confirmed defects from assumptions and checks that could not be run. Portfolio samples are original controlled demonstrations, not previous client engagements. Model training, live cloud integration, paid APIs and large repositories require separate scoping. One revision means correction or clarification within the agreed inputs and checks; new datasets or objectives are additional work.
Testing application:
Software
Device:
PC
FAQ
What is included in Basic?
One CSV up to 50,000 rows and 20 columns, one existing notebook or script up to 500 executable lines, up to 5 agreed checks, total inputs up to 10 MB.
What is included in Standard?
Up to 2 CSVs with 100,000 total rows and 30 columns each, up to 2 notebooks/scripts and 1,000 total executable lines, up to 10 agreed checks, total inputs up to 20 MB.
What is included in Premium?
Up to 3 CSVs with 250,000 total rows and 50 columns each, up to 3 notebooks/scripts and 1,500 total executable lines, up to 15 agreed checks, total inputs up to 30 MB.
Does this include training a new machine learning model?
The listed packages review an existing bounded analysis. New model development or training needs a separate agreed scope.
What if the workflow cannot be reproduced?
The report records the observed error, environment and missing prerequisites. Reproduction is not claimed when a required dependency or input is unavailable.
Are all findings automatically repaired?
The packages focus on validation and reporting. Any code repair must be explicitly included in the agreed scope.
How are AI tools used?
AI-assisted tools may support analysis and drafting. We will agree on any client restrictions and permitted data handling before work starts.
Can you guarantee model accuracy or publication acceptance?
No. Deliverables are reproducible checks and documented findings against an agreed scope, not guaranteed business or journal outcomes.
