Python for Industrial Automation: What It Can and Cannot Do

Python does not replace a PLC. It is not a real-time control language and it never will be. But the moment you stop thinking of Python as “another control language” and start thinking of it as the Swiss-Army knife that lives next to the PLC, the use cases explode: data analysis, report automation, PLC communication, log processing, anomaly detection with machine learning. If you work in automation in 2026, Python is no longer optional — it’s the second language every engineer should have.

Why Python in Automation

Python doesn’t compete with a Siemens CPU for the motor starter logic. It fills the gaps around the PLC — all the things a PLC cannot do, shouldn’t do, or is painful to do:

• Crunch thousands of rows of process data in seconds (Pandas)
• Generate shift reports in CSV / HTML / PDF without manual Excel labor
• Talk to the PLC over Modbus, OPC UA, or Snap7 from anywhere on the network
• Detect anomalies with machine learning and warn operators before a fault
• Push data to dashboards, databases, cloud or MES — pick your poison

Python is the number-one language in data science and machine learning globally. For an automation engineer that means access to the same tools data scientists use, applied to the data from your PLC, sensors, and production line.

It is also free, open-source, and runs everywhere. No matter the brand — Siemens, Mitsubishi, Wago, Allen-Bradley, Beckhoff — Python has a library that talks to it.

Practical Use Cases in the Plant

People ask: “OK, but what does Python actually do on the factory floor?” The honest answer is: everything the PLC either cannot or shouldn’t.

Production Data Analysis

Pull data from the PLC, import into Pandas, spot trends and anomalies. What takes hours of manual Excel work runs in seconds. A typical script: read 24 hours of temperature logs from the PLC, compute rolling averages, detect excursions beyond 3σ, flag the lot numbers produced during each excursion. This is the kind of analysis that catches a drifting valve two weeks before it fails.

Automated Reporting

Shift reports, production summaries, OEE dashboards. One Python script replaces an operator copying values from an HMI into an Excel template. Output formats: CSV for BI tools, HTML for internal dashboards, PDF for quality records.

Machine Diagnostics

Scripts that analyze motor currents, vibration, temperatures and generate readable diagnostic reports. Instead of reading raw values off a panel, you get a one-page summary: “Motor M12 current running 8% above baseline since last Thursday. Suggested action: inspect bearings.”

PLC Communication

Read and write to Siemens S7 via Snap7, to any Modbus device via pymodbus, to everything modern via OPC UA. Poll in real time, log the data, push it wherever it needs to go. See our Python + OPC UA tutorial for the S7-1500 case and our Python Modbus guide for the legacy-device case.

Browser Dashboards

Build HTML dashboards with charts, tables and KPIs without paying SCADA license fees. Plotly, Dash, Streamlit — any of them turn Python data into a browser dashboard in under an hour. For small plants and pilot projects, this is a complete alternative to traditional SCADA.

AI and Anomaly Detection

Predictive maintenance, process optimization, vision inspection. Python is the gateway to machine learning in automation. Libraries like scikit-learn, TensorFlow and PyTorch apply directly to sensor time-series data from your plant.

These aren’t hypotheticals — they’re real projects at Bosch, Nestlé, Lenze, Finder, and every automotive Tier-1 in Europe. Python complements traditional automation; it doesn’t replace it.

What You Actually Need to Know

You don’t need to be a full-stack developer. You need a solid foundation and a few industry-specific skills. Here’s the realistic learning path.

Python Basics

Start with the data types — int, float, str, bool — and the core structures: lists, tuples, dictionaries. Dictionaries are particularly useful in automation because they map naturally onto device states:

# Mapping device states with a dictionary
state = {
    "pump_1": "ON",
    "valve_3": "OFF",
    "motor_M2": "ALARM"
}

for device, status in state.items():
    print(f"{device} -> {status}")

Conditionals and Loops

if / elif / else is the same mental model as ladder logic in a PLC, just written differently. for and while loops iterate over sensor readings, alarm lists, or file rows. If you already program PLCs, you’ll pick these up in an afternoon.

Working with Files (TXT, CSV, JSON)

Industrial data ends up in files. Python reads, transforms and writes them effortlessly — from raw CSV to polished HTML reports. This single skill covers 50% of the daily scripting work of an automation engineer.

f-strings for Process Messages

f-strings are Python’s built-in template syntax. They make log messages and reports readable:

# Formatting process data with f-strings
temp = 78.5
pressure = 6.2
print(f"Temperature: {temp:.1f} °C | Pressure: {pressure:.2f} bar")
# Output: Temperature: 78.5 °C | Pressure: 6.20 bar

Functions and Modular Code

Wrap repeated logic in functions. A diagnose_motor() function you write once can run across dozens of motors in different projects. This is the point where your scripts stop being throwaway code and become reusable tooling.

Error Handling with try / except

Industrial environments are messy. Sensors fail, connections drop, files go missing. A production-grade script handles these with try / except so a single bad reading doesn’t crash the whole monitoring process.

Recommended order for learning: variables and types → conditionals → loops → lists and dicts → functions → file I/O. This path takes you from “Hello World” to a working industrial project in about 20-30 hours of focused practice.

VS Code or PyCharm?

One of the most common questions from beginners. The two dominant Python editors are Visual Studio Code from Microsoft and PyCharm from JetBrains. Both are powerful. For an automation engineer, one of them is clearly better.

Recommendation: VS Code

Use VS Code as your primary editor. It is free (entire product, no “Community” lobotomy), fast, lightweight, and wildly extensible.

Why VS Code in particular? First, it is fully free — no feature ceiling. Second, it launches in seconds where PyCharm takes 20-30 seconds. For an engineer who wants to open a quick diagnostic script, that speed matters. Third, VS Code is universal — today Python, tomorrow JSON configs, the day after XML from a PLC export. All handled with the same editor. PyCharm is Python-only, which is a real limitation in an automation workflow.

PyCharm’s one real advantage is its debugger and deep code intelligence on very large codebases. For automation work — scripts, integrations, data analysis — VS Code is more than enough, and its speed and universality win on daily work.

Essential VS Code Extensions for Automation

After installing VS Code, add these: Python (language support + debugger), Pylance (intelligent IntelliSense), Jupyter (notebooks for data experiments), GitLens (version control). With that stack, you have a professional engineering environment ready to go.

Getting Started — Practical First Steps

The best way to learn is with real automation problems, not abstract exercises. Skip the “sum numbers from 1 to 100” kata — write your first temperature alarm script instead.

Step 1: Install Python from python.org and VS Code with the Python extension. Write your first program — a classic print("Hello from Python"). Confirm the environment works before moving on.

Step 2: Learn the conventions early. Follow PEP 8. Give variables descriptive names — furnace_temperature instead of t, system_pressure instead of p. Code you can read six months later is code you can maintain.

Step 3: Solve small automation problems. Display a startup banner for a machine. Check whether temperature and pressure are within range. Iterate over a list of sensors and print their statuses. Each of these connects Python to a real industrial context from day one.

Step 4: Work with data. Read process CSVs, filter them, spot anomalies, generate summaries. This is the skill that converts directly into workplace value.

Project: The Factory Data Analyzer

The most effective way to learn is building a complete project end-to-end. A proven template: a Factory Data Analyzer — an application that pulls data from the process, logs it, simulates scenarios, and produces CSV + HTML reports.

Such a project touches every essential skill:

• Motor diagnostics — analyze current, temperature, rpm, alarms with if/else, generate a per-motor status report.
• Logging and monitoring — write time-series data to files, monitor PLC tags in a loop, alarm logs in CSV and TXT.
• Startup banners — practical use of f-strings and conditionals to generate operator messages.
• PLC tag monitoring — continuous sensor scan in a while loop with proper exception handling.
• CSV / TXT logging — persistent storage of process data for later analysis.
• HTML production report — final visualization with KPIs, trend charts, data tables rendered in the browser.

Bonus: connect to a real PLC (Siemens S7-1200/1500) via Snap7 or Modbus. If you don’t have hardware, use a simulator. The visualization layer turns into an interactive browser dashboard — no expensive SCADA license required.

Learning Python in Context

Self-teaching Python is doable — the docs are great, the community is enormous. The problem is that 95% of the material uses web-development examples that feel foreign to an automation engineer. You learn loops by iterating over shopping carts, not sensor readings. You learn classes by building a blog post, not a motor diagnostic.

The fix is to learn Python in your context from the start. Variables that are furnace_temperature and system_pressure, not name and age. Projects that analyze PLC data and produce shift reports, not todo lists.

FAQ

Does Python run on a PLC?

Not directly on a traditional Siemens or Allen-Bradley CPU. Some modern edge platforms (Siemens IoT2050, Beckhoff TwinCAT BSD, Phoenix Contact PLCnext) run Python natively as an edge layer next to the control runtime. For everything else, Python lives on a PC, laptop, or Raspberry Pi on the same network as the PLC.

Can I use Python for real-time motion control?

No. Python is not real-time. Use it for supervisory logic, data acquisition and reporting — always above the control layer, never below it. Motion and safety logic belong in the PLC or dedicated drives.

Which Python libraries should an automation engineer learn first?

pandas for data, pymodbus for Modbus, asyncua for OPC UA, python-snap7 for Siemens S7Comm, matplotlib or plotly for charts, requests for HTTP APIs. That six-pack covers 80% of real-world automation scripting.

Is Python fast enough to process data from a high-speed line?

For polling rates down to 10 ms and data rates up to millions of rows per day, yes, easily. For sub-millisecond or microsecond work, drop to C extensions or Cython — but at that rate you should probably be doing the work in the PLC anyway.

Does Python replace SCADA?

For small installations and pilot dashboards — yes, often it does. Plotly Dash and Streamlit build perfectly usable browser dashboards in hours. For site-wide SCADA with alarm management, operator authentication, audit trails and redundancy, a proper SCADA platform (Ignition, WinCC, FactoryTalk View SE) is still the right answer.

Should I learn Python 2 or Python 3?

Python 3. Python 2 has been end-of-life since 2020. All modern libraries (pymodbus, asyncua, pandas) are Python 3 only.

What’s Next

Python for industrial automation is a skill multiplier. Combined with PLC knowledge, it opens doors from engineering into data, from data into ML, and from ML back into process optimization. The gap between engineers who use Python daily and those who don’t is widening.

If you want to learn Python specifically in an automation context — with pandas, Modbus, Snap7, HTML reports and a full Factory Data Analyzer project — our Python for Automation Engineers course runs through all of it in Visual Studio Code, from the first print() to deploying dashboards next to a real PLC.

For the communication layer (Modbus, Profinet, OPC UA side by side), see Communication Protocols. For the full Siemens engineering stack, Siemens PLC Training gets you from basic tags to OOP in V20.

Author

Simon Adamek

IT Engineer and PLC Specialist
Manager at ControlByte
“Guiding beginners into the world of PLCs and industrial innovation.”

See author's posts