virtualcommons / port-of-mars-analysis Goto Github PK

https://colab.research.google.com/

https://towardsdatascience.com/a-beginners-guide-to-sentiment-analysis-in-python-95e354ea84f6

available time blocks for the player should also be listed in players.csv alongside how much they invested (due to changes from events / personal gain / voted hero in hero or pariah)

The handshake between https://github.com/virtualcommons/port-of-mars/blob/main/dump.sh and running the routines here is unnecessarily complicated by the input data requirements.

Refactor either the dump scripts or the import routines so that we can feed the output from dump.sh directly into this repo.

For cultural, pre and post surveys

• rows missing a participant_id or invite_id should be outside the tournament date range

Our current dataset satisfies this property but maybe not a tournament in the future

we can draw relationships between language (e.g. frequency using words like pariah, hero, etc) and severity of events

need to define what severity index encompasses for an categorizing an event

Treat Google Shared Drive as a staging area

Need a place to publish the actual data (privately at first / embargoed until publication)

Data needs clear documentation (where things are, what they mean, how they were generated / provenance, etc)

OSF https://osf.io/
FigShare https://figshare.com
Zenodo https://zenodo.org

A data collection diagram like the following can be useful for documenting how we get from experiment data -> research findings:

Moosavi, Sobhan, Mohammad Hossein Samavatian, Srinivasan Parthasarathy, Radu Teodorescu, and Rajiv Ramnath. “Accident Risk Prediction based on Heterogeneous Sparse Data: New Dataset and Insights.” In proceedings of the 27th ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems, ACM, 2019.

There is an issue with the analysis pipeline integrating round 1 pre surveys with round 2+ pre surveys. Not sure if this is a data issue where the round 1 pre survey has changed and the round 2+ pre surveys did not or vice versa.

@MarcoAJanssen

<error/rlang_error>
Error in `dplyr::union_all()` at R/survey.R:301:2:
! `x` and `y` are not compatible.
✖ Different number of columns: 111 vs 122.
---
Backtrace:
    ▆
 1. └─global tournament_load(tournament_dir, max_game_rounds)
 2.   └─global survey_tournament_load(...) at R/tournament.R:7:2
 3.     ├─dplyr::union_all(r1, r2plus) at R/survey.R:301:2
 4.     └─dplyr:::union_all.data.frame(r1, r2plus)

1. Personal Gain (VotedForPersonalGain)
2. Hero or Pariah (VoteHeroOrPariah)
3. Compulsive Philanthropy (VotedForPhilanthropist)
4. Breakdown of Trust: keep up to 2 influence resource units (BreakdownOfTrust
5. Bonding through Adversity
6. Efforts Wasted
7. any others?

Identify a way to programmatically discover these events and emit their data payloads in MarsEventSummarizer, possibly GameEventWithData will work?

We should place all the input data needed to regenerate the processed / integrated data (survey + actions) into the shared google drive along with the processed / integrated data

The data file that includes information on events that happened in the round has a number of issues:

• Sandstorm is listed for 2 rounds while a sandstorm should be happening for 3 rounds. Is this an accounting error?
• In game 64 there are no events for round 8, which is not possible
• In game 69 there are no events for rounds 2 and 3
• It would be great to know the voting outcomes, like whether hero or pariah is elected and who the hero or pariah is.

add documentation and structure to R scripts

use RStudio ==== to demarcate sections in the script and document its organization + control flow better