EV Car Data Scraping Updates 11/8
Alex Liu, Sally Yin, Rob Carleski, Sabina Tomkins
Table of contents
Summary of work
Engineering challenges- getting all listings
Sample datasets- EV and All cars
Next steps
Summary of work
Pipeline development to get all car listings
Improved metric reporting
Dataset preparation, summary statistics
Engineering challenges
scheduled daily at
8:00 am
send
HTTP request
static website
dynamic website
car sales web server
get web content on eligible car sales listings
parse content into individual listing urls
car sales web server
car sales web server
car sales web server
car sales web server
parse listing information from HTML into JSON
get web content on individual listing
send
HTTP request
send
HTTP request
send
HTTP request
send
HTTP request
Engineering challenges
scheduled daily at
8:00 am
send
HTTP request
static website
dynamic website
car sales web server
get web content on eligible car sales listings
parse content into individual listing urls
car sales web server
car sales web server
car sales web server
car sales web server
parse listing information from HTML into JSON
get web content on individual listing
send
HTTP request
send
HTTP request
send
HTTP request
send
HTTP request
EV's: 10's of requests
All cars: 1,000's
Engineering challenges
scheduled daily at
8:00 am
send
HTTP request
static website
dynamic website
car sales web server
get web content on eligible car sales listings
parse content into individual listing urls
car sales web server
car sales web server
car sales web server
car sales web server
parse listing information from HTML into JSON
get web content on individual listing
send
HTTP request
send
HTTP request
send
HTTP request
send
HTTP request
EV's: 10's of requests
All cars: 1,000's
Engineering challenges
scheduled daily at
8:00 am
send
HTTP request
static website
dynamic website
car sales web server
get web content on eligible car sales listings
parse content into individual listing urls
car sales web server
car sales web server
car sales web server
car sales web server
parse listing information from HTML into JSON
get web content on individual listing
send
HTTP request
send
HTTP request
send
HTTP request
send
HTTP request
EV's: 10's of requests
All cars: 1,000's
EV's: 1,000's of listings; All cars: 100,000's
Engineering challenges
scheduled daily at
8:00 am
send
HTTP request
static website
dynamic website
car sales web server
get web content on eligible car sales listings
parse content into individual listing urls
car sales web server
car sales web server
car sales web server
car sales web server
parse listing information from HTML into JSON
get web content on individual listing
send
HTTP request
send
HTTP request
send
HTTP request
send
HTTP request
EV's: 10's of requests
All cars: 1,000's
EV's: 1,000's of listings; All cars: 100,000's
BLOCKED
SLOW
Engineering challenges- getting all listings
Sample datasets- Configuration
Vehicle type | EV | All cars |
Dates | 10/27-11/2 | 11/6-11/7 |
Sample datasets- Configuration
Vehicle type | EV | All cars |
Dates | 10/27-11/2 | 11/6-11/7 |
Search ranges | 17 cities, 50mi | 17 cities, 50mi |
Platform | Platform 1 | Platform 1 |
Sample datasets- Web scraping success
Dataset | EV | All cars |
Listings identified | 11,795 | 174,132 |
Scraped listings | 10,761 | 169,810 |
Unique listings | 2,774 | 156,538 |
Sample datasets- Data fields success
Dataset | EV | All cars |
Unique listings w/ Essential columns | 100% | 100% |
Ess. + Important | 83% | 78% |
Ess. + Imp. + Nice | 45% | 47% |
Sample datasets- Basic statistics (price)
Vehicle type | EV | All cars |
Mean price | $30,260 | $17,733 |
Max price | $448,888 | $999,999 (max) |
Sample datasets- Basic statistics (price)
Sample datasets- Basic statistics (price)
EV's
All cars
Sample datasets- Basic statistics (proportion of pop.)
EV's
All cars
Sample datasets- Basic statistics (EV:Car ratio)
Sample datasets- Basic statistics (fuel type)
Next steps
Share these sample datasets
Build out more robust pipeline for collecting all cars
Investigate data loss for systematic bias
Questions