Session 4: Applied AI

1 Introduction

In this final session, the focus shifts to you. This is a hands-on session designed for you to apply the AI tools and techniques learned throughout the day to your own transport problems and datasets.

2 Objectives

Apply AI Skills: Use AI coding assistants, data analysis tools, and LLMs to solve real-world challenges relevant to your work.
Problem Solving: Tackle a specific problem or question you brought with you, or explore a provided dataset in depth.
Peer Learning: Share approaches and insights with colleagues, fostering a collaborative learning environment.

3 Session Structure

3.1 Problem Definition (15 mins)

Briefly define the problem you want to solve or the question you want to answer.
Identify the data you have available or need to acquire.
Determine which AI tools (e.g., ChatGPT, Claude, GitHub Copilot) are most suitable for your task.

3.2 Hands-on Development (60 mins)

Work on your project: Spend this time coding, analyzing data, or writing reports with the assistance of AI.
Instructor Support: The instructors will be circulating to provide one-on-one guidance and troubleshoot issues.
Collaboration: Feel free to work in pairs or small groups if you have similar interests or datasets.

3.3 Show and Tell (15 mins)

Volunteers will have the opportunity to briefly share what they have achieved or discovered.
Discuss challenges faced and how AI helped (or didn’t help) overcome them.

4 Data for the Session

If you haven’t brought your own data, you can use the datasets prepared for this course. They are available on the Releases page of the repository. The files have been downloaded into the working directory.

library(sf)

# TfSE Boundary and Geographies
tfse_boundary = read_sf("tfse_boundary.geojson")
tfse_lads = read_sf("tfse_lads.gpkg")
tfse_msoas = read_sf("tfse_msoas.gpkg")

# Motorway Network
motorways = read_sf("south_east_motorways.gpkg")

# Collision Data
collisions = read_sf("collisions_tfse_2020_2024.gpkg")

# OD Data
od = read.csv("od_tfse_2021.csv")

import geopandas as gpd

# TfSE Boundary and Geographies
tfse_boundary = gpd.read_file("tfse_boundary.geojson")
tfse_lads = gpd.read_file("tfse_lads.gpkg")
tfse_msoas = gpd.read_file("tfse_msoas.gpkg")

# Motorway Network
motorways = gpd.read_file("south_east_motorways.gpkg")

# Collision Data
collisions = gpd.read_file("collisions_tfse_2020_2024.gpkg")

# OD Data
import pandas as pd
od = pd.read_csv("od_tfse_2021.csv")

5 Suggested Activities

If you didn’t bring your own data, here are some challenges you can try using the open datasets we’ve looked at today:

Advanced Visualization: Create a dashboard summarizing the OD data we used in Session 2.
Policy Analysis: Use an LLM to compare two different transport policy documents and identify conflicting recommendations.
Automated Reporting: Write a script that automatically generates a summary report from a CSV file of traffic counts.