Turn raw CPT sounding files into one consistent workbook and profile plots with an AI-written script

Short answer

Use OpenAI Codex, Claude Code, ChatGPT or Claude: The AI writes a Python script that reads each CPT file layout into one standard table, applies only the equations you pasted from your own reference, tests itself against your hand-calculated rows, logs unparsed data, and writes a workbook and fixed-scale profile plots. CPT data arrives in several file layouts (comma or space separated, different header blocks, different units), and reformatting each sounding before you can look at it eats time. Have Codex, Claude Code, ChatGPT or Claude write a short Python script that reads each layout into one standard table, applies only the equations you paste from your own reference, checks itself against a small hand-calculated sounding, and writes a workbook and consistent profile plots. Practitioners are rightly wary of home-built CPT tools, so the script is validated against your existing CPT software before anyone relies on it, and interpretation stays with the engineer. A person still checks: validation against an independent CPT tool, every warning in the processing log, and all layering, correlations, design parameters and liquefaction judgments.

AI tie-in

Tool: OpenAI Codex, Claude Code, ChatGPT or Claude

The AI writes a Python script that reads each CPT file layout into one standard table, applies only the equations you pasted from your own reference, tests itself against your hand-calculated rows, logs unparsed data, and writes a workbook and fixed-scale profile plots.

Prompt 1

Here are the first 30 lines of three CPT data file layouts we receive [paste, with project details removed]. Write a commented Python script that reads all three into one table with columns: sounding ID, depth, cone resistance, sleeve friction, pore pressure, each in the units I specify here [units], plus the header values cone area ratio, predrill depth and groundwater depth. Then apply only the equations in my config file [paste equations with units] and do not use any equation, coefficient or chart boundary from memory. Log every row or header you cannot parse instead of guessing, write an Excel workbook with one tab per sounding, save profile plots at fixed scales I choose, and include a test that reproduces my hand-calculated rows [paste].

Prompt 2

Here is the script's output for one sounding [paste] and the same sounding processed in our existing CPT software [paste]. Compare them depth by depth, list every difference larger than [tolerance], and for each one point to the line of code or the input setting (units, area ratio, depth reference, groundwater level) that could explain it. Do not change the equations; suggest fixes only for coding or input errors.

Human check: Validation against an independent CPT tool, every warning in the processing log, and all layering, correlations, design parameters and liquefaction judgments.

The problem

Turn raw CPT sounding files into one consistent workbook and profile plots with an AI-written script. This is a common, time-consuming task for a geotechnical engineer.

What this workflow does

CPT data arrives in several file layouts (comma or space separated, different header blocks, different units), and reformatting each sounding before you can look at it eats time. Have Codex, Claude Code, ChatGPT or Claude write a short Python script that reads each layout into one standard table, applies only the equations you paste from your own reference, checks itself against a small hand-calculated sounding, and writes a workbook and consistent profile plots. Practitioners are rightly wary of home-built CPT tools, so the script is validated against your existing CPT software before anyone relies on it, and interpretation stays with the engineer.

How it works: AI drafts or analyzes. A person operates the tools, checks the work, and takes action.

Your checkpoint: The script reformats data and does the arithmetic you defined; it does not interpret soil behavior. Validate it against an independent CPT tool, review every warning, and leave layering, correlations, design parameters and liquefaction judgments to the geotechnical engineer.

Before you start

Access to the named tool; confirm license, version and cost with the vendor.
A small, non-confidential sample to try the workflow on first.
A qualified person to review the output before it is used.

Tools and access

Follow the steps

Collect one sample file of each layout you receive and note what the header tells you: units, depth reference, cone area ratio, predrill depth and groundwater level used. Strip client and project details from the samples before sharing them with any AI tool, per your firm's policy.
Paste the first 30 or so lines of each layout into Codex, Claude Code, ChatGPT or Claude and have it write a reader that maps every layout to the same columns (sounding ID, depth, cone resistance, sleeve friction, pore pressure, plus header values), converts units the way you tell it, and logs any row or header it cannot parse instead of filling it in.
Type the correction and normalization equations you use, with units, into a small config file from your own licensed reference, along with any classification boundaries. Tell the model to implement exactly those and not to supply equations, coefficients or chart boundaries from memory.
Have it write tests before you trust it: a five-to-ten-row sounding you calculate by hand that the script must reproduce, plus data checks for depth that does not increase, zero or negative readings, spikes and values outside the range you set.
Output one workbook with a tab per sounding and a summary tab, profile plots at the same scales for every sounding, and a processing log listing each file, the settings used and every warning.
Validate on one or two soundings already processed in your existing CPT software and explain every difference. Layering, parameter selection, correlations and any liquefaction screening remain engineering judgment. Save the script version with each job.

Try a small example

Replace the bracketed fields with your own project details. No tested sample run is published for this workflow yet.

Sample input

Fill in the bracketed [fields] in the text below with your project data.

Starter prompt

Here are the first 30 lines of three CPT data file layouts we receive [paste, with project details removed]. Write a commented Python script that reads all three into one table with columns: sounding ID, depth, cone resistance, sleeve friction, pore pressure, each in the units I specify here [units], plus the header values cone area ratio, predrill depth and groundwater depth. Then apply only the equations in my config file [paste equations with units] and do not use any equation, coefficient or chart boundary from memory. Log every row or header you cannot parse instead of guessing, write an Excel workbook with one tab per sounding, save profile plots at fixed scales I choose, and include a test that reproduces my hand-calculated rows [paste].

Expected output

Not published yet. Treat any output as a draft and check it against the source documents.

Check the result

Every value, citation or quantity traces back to a source document.
Missing information is marked as missing, not filled in.

Limits and recovery

The script reformats data and does the arithmetic you defined; it does not interpret soil behavior. Validate it against an independent CPT tool, review every warning, and leave layering, correlations, design parameters and liquefaction judgments to the geotechnical engineer.
Check your organization’s data policy before sharing project files with third-party tools.

Stop or switch when: The source cannot be verified, permissions are missing, or the result contains unsupported facts.

Frequently asked questions

How does AI help with this?

OpenAI Codex, Claude Code, ChatGPT or Claude: The AI writes a Python script that reads each CPT file layout into one standard table, applies only the equations you pasted from your own reference, tests itself against your hand-calculated rows, logs unparsed data, and writes a workbook and fixed-scale profile plots. A person still checks: Validation against an independent CPT tool, every warning in the processing log, and all layering, correlations, design parameters and liquefaction judgments.

How do I turn raw CPT sounding files into one consistent workbook and profile plots with an AI-written script?

Collect one sample file of each layout you receive and note what the header tells you: units, depth reference, cone area ratio, predrill depth and groundwater level used. Strip client and project details from the samples before sharing them with any AI tool, per your firm's policy. Paste the first 30 or so lines of each layout into Codex, Claude Code, ChatGPT or Claude and have it write a reader that maps every layout to the same columns (sounding ID, depth, cone resistance, sleeve friction, pore pressure, plus header values), converts units the way you tell it, and logs any row or header it cannot parse instead of filling it in. Type the correction and normalization equations you use, with units, into a small config file from your own licensed reference, along with any classification boundaries. Tell the model to implement exactly those and not to supply equations, coefficients or chart boundaries from memory. The full step list is above.

Does this use an AI agent or regular software?

It is labeled AI assists: AI drafts or analyzes. A person operates the tools, checks the work, and takes action.

What tools do I need?

OpenAI Codex, Claude Code, ChatGPT, Claude, Python, Microsoft Excel. Confirm license, version and cost with the vendor, and try it on a small non-confidential sample first.

Has this workflow been tested?

Workflow guide · not independently tested. Written from vendor documentation and practitioner use. Labels apply to the steps described here, not every feature of a product. Links were checked when the page was written.

Evidence and maintenance

Written from vendor documentation and practitioner use. Labels apply to the steps described here, not every feature of a product. Links were checked when the page was written.