data driven skate stuff for 4plymag
Example dataset to follow: andrew allen data
The shittiest part about this is creating the data. But it's really not that bad. Here's how:
We want a dataset with the following schema:
| Column | Description | Data Type |
|---|---|---|
clip_index |
The number of the video clip in which the clip occurred. For example, if the third clip of a video part is a line, each clip in the line should have a clip_index of 3. |
text |
trick_index |
The trick number, i.e. the ith trick in the video should have a trick_index of i. Each trick_index should be unique and consecutive, unless the trick is shown multiple times (e.g. double-angle or slow motion). |
text |
trick |
Name of the trick. Try to use 'fs' and 'bs', and separate tricks in manuals with 'to'. | text |
switch |
Boolean indicating whether or not the trick was switch. 1 if yes, else leave blank. |
number |
line |
Boolean indicating whether or not the trick was done in a line. 1 if yes, else leave blank. |
number |
slowmo |
Boolean indicating whether or not the trick was filmed in slow motion. 1 if yes, else leave blank. |
number |
obstacle |
Not-very-detailed description of obstacle. Should be rail, ledge, manual, gap, stair, flat, hip. (Use obstacle_detailed column to add more description). |
text |
obstacle_detailed |
A slightly more detailed description of the obstacle. Try to use what's been used before (e.g. for rail use handrail or shootout or flatbar, or for ledge, use hubba, etc.). |
text |
location |
Location of trick, if it's notable. Otherwise leave blank (don't worry too much about this). | text |
video |
Name of video. | text |
That is, for a given video part, add a row to the dataset above for each trick.
If you think of something interesting, feel free to add it as a column. For example, I noticed that Greco skates a ton of banks/slopes/etc., so I added a column called has_bank to his dataset.
For Koston's part in the 101 promo, I created this dataset.
For all of Greco's parts, I created this dataset.
Once you're done, save the data as a csv or xlsx document.
Then:
You Know Git: Fork the repo, add data, and make a pull-request.
You Don't Know git: Email me the data/description! jdwlbr at gmail