Module 2: Introduction to low-code Python programming
AI-aided content analysis of sustainability communication
Lesson 2.2: Python and AI-aided data analysis
lecture slides
lecture video
lecture text
AI-Aided Quantitative Analysis with Python
Artificial intelligence—especially large language models—now supports researchers in performing quantitative analysis more efficiently. These models can generate Python code on demand for statistical tasks, allowing users to interact with their data through natural language. This reduces the technical barrier for non-programmers while enabling rapid hypothesis testing, data cleaning, and model generation. The balance between manual coding and AI-assisted scripting gives researchers more control and efficiency, particularly when iterating through complex analytical workflows.
AI-Aided Computational Content Analysis
AI is also transforming how we approach content analysis. Large language models can assist in extracting meaning from text through classification, summarization, or topic modeling. Meanwhile, vision models like CLIP or ViT enable analysis of image and video content. Together, these tools support multimodal analysis, where textual and visual data are examined in tandem. Researchers can automate tasks such as entity recognition and theme detection, significantly speeding up the analysis of large, complex corpora while still allowing for nuanced interpretation.
Low-Code Data Analysis with Tables
Python makes it easy to read and manipulate tabular data, especially using the pandas library. With AI-assisted tools, users can perform complex filtering, grouping, and aggregation with simple prompts rather than verbose code. This low-code approach is ideal for exploratory data analysis, especially in non-technical fields. Tools like pandasgui, datatable, or conversational agents like ChatGPT can generate, execute, and explain table operations in real-time. This democratizes access to data insights and facilitates reproducibility.
Data Transformations with AI Assistance
Transforming data—through selection, filtering, and aggregation—is foundational to any analysis workflow. In Python, these operations are traditionally handled with pandas, but AI tools can now suggest or execute them via natural language instructions. This makes it easier for users to identify the variables they need, set conditions for filtering, and compute summaries like group means or totals. By comparing AI-generated queries to traditional code, learners gain a deeper understanding of the logic behind data manipulation.
Low-Code Data Analysis with Graphs
Visualization is a key part of understanding and communicating data. Python supports a wide range of basic plotting options, including bar charts, line graphs, and histograms. With AI assistance, users can describe the kind of graph they want and receive ready-to-run code. This enables quick generation of univariate and bivariate plots that explore single variables or pairwise relationships. Libraries such as matplotlib, seaborn, and plotnine provide powerful tools for crafting clear, insightful visualizations with minimal setup.
Multivariate and Interactive Graphs
When analysis involves more than two variables, visualization becomes more complex—and more powerful. Python offers tools to create multivariate plots like scatter matrices, heatmaps, and bubble charts. For dynamic exploration, libraries like plotly, altair, and holoviews enable interactive features such as zooming, filtering, and tooltip displays. AI support can refine these visualizations by adjusting parameters or recommending plot types based on data structure. This helps users engage deeply with high-dimensional data and uncover hidden patterns.