LEMONGRAB

Gather company data from mobygames and wikidata and build awesome research datasets and visualizations.This tool provides various commands to fetch data (via cli) and visualize it (via browser).

Requirements

• Python 3.7
• unified api
• diggrtoolbox

Prerequisites

It is recommended to install lemongrab in a virtual Python environment such as Pipenv, virtualenv, or venv.

Installation

Clone this repository and install the package.

$ git clone https://git.sc.uni-leipzig.de/ubl/diggr/general/lemongrab
$ cd lemongrab

Open lemongrab/settings.py with an editor of your choice and edit the value of DIGGR_API to the address of your instance of the UnifiedAPI. Save the file and install lemongrab.

$ pip install .

Initial setup / Create a project

Create a folder and initialize lemongrab.

$ mkdir testproject && cd testproject
$ lemongrab init

This will create two directories, one for the company_networks and on one for the required datasets. Next: Create both the wikidata mapping and the mobygames companies dataset by running:

$ lemongrab build all

Note: If you already build the datasets somewhere else, you can copy those files into the lemongrab_datasets directory and save yourself some time.

Features

The tool provides two main commands company-network and browser. The first command builds a GraphML file to be inspected (e.g. with Gephi). The second command starts a web application which can be used to inspect the companies fetched earlier.

$ lemongrab company-network -c "Japan" -p "Sony PlayStation" --roles
$ lemongrab browser

Every company network is supplied with a log file. To aggreate all logs of your project simply invoke the aggregate-logs command.

$ lemongrab aggregate-logs

This will build a CSV file in your project directory with the contents of all log files in your current project.

Usage

company-network

Builds a network of companies based on their common work on games

Option	Description	Multiple	Default value
--country/-c	Release country	X	--
--platform/-p	Release platform	X	--
--gamelist/-g	Use tulpa gamelist as filter	--	--
--roles/--no-roles	Differentiate companies by their production roles	--	--no-roles
--publisher/--no-publisher	Include/exclude publisher roles; only needed when --roles option is set	--	--no-publisher

You can either filter by Country/Platform OR tulpa gamelist.

Options which allow multiple invokation (currently country and platform) can be used multiple times in the same call, e.g.:

$ lemongrab company-network -c Japan -c Worldwide

game-company-sample-network

Builds a network of companies based on a tulpa generated company sample.

$ lemongrab game-company-sample-network ../tulpaproject/datasets/tulpa-companies.json

browser

Opens the lemongrab browser frontend for data exploration

Datasets

Mobygames companies dataset

lemongrab_datasets/mobygames_companies.json

This dataset contains all companies and the games they were working on.

Structure:

{
    "<comapny_id>": [
        {
            "company_name": "Fox Interactive, Inc.",
            "game_id": "672",
            "game_slug": "die-hard-trilogy",
            "game_title": "Die Hard Trilogy",
            "game_years": [
                1996,
                1997,
                1998
            ],
            "production_role": "Published by",
            "release_countries": [
                "Sweden",
                "United Kingdom",
                "Italy"
            ],
            "platform": "Sony PlayStation"            
        },
    ]
}

Wikidata mapping and country information

lemongrab_datasets/wikidata_mapping.json

Contains a mapping of Mobygames company slugs to wikidata items as well as country information from wikidata (if available)

Structure:

[
    {
        "mobygames_slug": "company-slug",
        "country": "Japan",
        "wkp": "Q111111"
    },
]

Mobygames company id to slug mapping

datasets/mobygames_companies_id_to_slug.json

Mapping of the internal Mobygames company ids to their slug

Structure:

[
    {
        "company_id": "<company id>",
        "slug": "company slug"
    },
]

License

GPLv3

Authors

• Peter Mühleder [email protected]
• Florian Rämisch [email protected]

Copyright

2019-2020, Universitätsbibliothek Leipzig, [email protected]