InnovationNetworks

Repository for the Innovation Networks project, accepted for NetSciX 2024.

About

This project takes fits empirical social netowrks to an agent-based model (ABM). After determining the reinforcement and novelty parameters, the ABM generates multiple 50-node networks to represent the original network. Innovation simulations are then conducted to assess how reinforcement and novelty impact innovation efficiency. The project also compares different empirical networks with innovation-optimal networks to identify differences.

Table of Contents

• Installation
• Usage

Usage

1. Calculate the network metrics for your social network by running the code in calculate_metrics. This will create data/metrics/<dataset>.csv with the metrics of the empirical network.

2. Fit your network to the ABM by running the code in fitting. This will create fitting/results/<dataset>/best.csv with the best fit reinforcement and novelty parameters using QD algorithms and the metrics from step 1.

3. Run the ABM and calculate the resulting innovation efficiencies by running the code in empirical. This will create empirical/results/<dataset>/output.csv with the mean NCTF and TTF of the ABM0-generated networks.

4. Run the full search to running the code in full_search. This will create full_search/results/<innovation_type>/output.csv with the mean NCTF and TTF of the ABM-generated networks for all possible reinforcement and novelty parameters.

5. To visualise the results in the browser, the optimal and least optimal innovation networks must be converted into JSON format. This is done by running full_search/analyse.py. This will create 100 networks for the best and worst choice of ABM parameters and create correspodning JSON files in full_search/data/output/graph/.

6. To run the webapp navigate to webapp and run npm install and npm start. This will start the webapp on localhost:3000. The webapp can be used to visualise the networks in the browser. You need to ensure the correct JSON files are placed in webapp/src/data/. See the webapp README for more details.

7. To create plots of the results, run the scripts in visualise. This will create plots in visualise/results/.