Crunchbase Data

From edegan.com
Revision as of 16:10, 6 September 2017 by Peterjalbert (talk | contribs)
Jump to navigation Jump to search


McNair Project
Crunchbase Data
Project logo 02.png
Project Information
Project Title Crunchbase Data
Owner Adrian Smart
Start Date June 2017
Deadline
Keywords Data, Tool, Crunchbase, VC, Angel
Primary Billing
Notes
Has project status Active
Copyright © 2016 edegan.com. All Rights Reserved.


Files and dbases

The dbase is:

crunchbasebulk

The files are in:

E:\McNair\Software\Database Scripts\Crunchbase
Z:\Crunchbase\CrunchbaseData

To do

  1. Download the data
  2. Extract top 5 lines
  3. Build table specs
  4. Import the data

Issue importing events.csv. Line count does not match import count. Also typos in data for start_time and end_time which created import errors with time datatype. Switched to varchar. Issue importing organizations.csv. Line count is 515496 but on import got 515491. Last 4 lines in file are chinese characters which might affect import.

Downloading the data

Our user key is: 662e263576fe3e4ea5991edfbcfb9883

  1. The download script is written in perl. It is called downloadScript and located in E:\McNair\Software\Database Scripts\Crunchbase. You can execute it by typing "perl downloadScript.pl" in terminal.


Where, what, etc.

Importing the data

To import the data make sure that all 22 crunchbase csv files are on the db server in /bulk/crunchbase/crunchbaseData. Also make sure that the load_crunchbase.sql script is in this directory. Run psql crunchbasebulk to start the db from this directory. Run the command "\i load_crunchbase.sql" to run the script. This will load the contents of the 22 csv files into the db. Check that the number of lines copied into each table matches the actual lines in the csv file. The line numbers have been included in the comments of the load_crunchbase.sql script. See issues section for unexpected results.

Type \dt to get a list of the tables in the crunchbasebulk db. Type \q to quit db. When the db is not running you can type 'wc -l acquisitions.csv' to get a line count of this file.


News is a relationship query parameter

https://api.crunchbase.com/v/3/organizations?user_key=662e263576fe3e4ea5991edfbcfb9883&uuid=1e4f199c-363b-451b-a164-f94571075ee5

permalink is airbnb in this example:

https://api.crunchbase.com/v/3/organizations/airbnb/news?user_key=662e263576fe3e4ea5991edfbcfb9883

Same thing for people - people end point then permalink

Accelerator Founders Data

The Crunchbase API can be used to readily access Founders Data. The API is used in the web format: https://api.crunchbase.com/v/3/organizations/company_name/?user_key=662e263576fe3e4ea5991edfbcfb9883, to get a JSON object with alot of data.

One such field in the JSON object is Founders, which is followed by profiles of the founders for any given company. Not all the accelerators have a crunchbase page, but it is a good start.

I queried the API for all the accelerators we have listed in the following file:

E:\McNair\Projects\Accelerators\accelerators.txt

I retrieved a list of founders for the accelerators that returned results from Crunchbase. Out of the 269 accelerators we have on record, 136 of them turned up results for founders, resulting in 312 founders. The list of founders and their respective company can be found at:

E:\McNair\Projects\Accelerators\founder_names.txt

A table has also been added to the crunchbasebulk database called founders. The table contains 3 columns: Company, first_name, last_name. Another table, called founders_linkedin, contains 4 columns: Company, first_name, last_name, and linkedin_url.

The file has been exported to the McNair DB Server.

The next step is to match each founder name with a linkedin profile. These profiles can be accessed using our LinkedIn Crawler to gather more information about each founder.