phenrickson.github.io/boardgames.qmd at main · phenrickson/phenrickson.github.io · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
---
title: "Board Games"
---

---

If you love boardgames and data, [**boardgamegeek.com**](https://boardgamegeek.com) is a treasure chest of data on every boardgame in existence. What can we learn from this data?

## [Predicting Upcoming Board Games](https://phenrickson.github.io/bgg_models/)

<center>

![](images/bgg_outcomes.png){width=50%} ![](images/bgg_predictions.png){width=40%}

</center>

What upcoming board games are **likely to be popular within the BGG community?**

I train a variety of predictive models on historical BGG data in order to estimate a variety of outcomes for newly released games. Why are some games expected to better than others? How accurately can we predict upcoming games? What games are likely to become the new hotness? Take a look to see what the models think.

[Predictions](https://phenrickson.github.io/bgg_models/)

## [Predicting Board Game Collections](https://phenrickson.github.io/bgg_collections/)

![](images/phenrickson_coefs.png){width=40%} ![](images/phenrickson_sep.png){width=50%}

"What upcoming games should I check out?"

I train models at the individual user level to predict whether someone will be likely to add an upcoming game to their collection. This started as a project to [train a model to identify games for myself](https://phenrickson.github.io/bgg_collections/phenrickson.html), but then I realized that this could be extended to anyone with a boardgamegeek account and a decent number of games in their collection.

[Collection Models](https://phenrickson.github.io/bgg_models/)

## [Collecting BGG Data](https://phenrickson.github.io/bgg_data/)

![](images/bgg_data_pipeline.png){width=40%} ![](images/bgg_data_batch.png){width=50%}

Using the [targets](https://github.com/ropensci/targets) package to build a pipeline with dynamic branching for weekly API requests from BGG's API.

## [Adjusting BGG Ratings: Complexity and 'The Hotness'](https://phenrickson.github.io/bgg_reports/reports/adjusting_bgg_ratings.html)

<center>

![](images/averages_and_usersrated.png){width=45%} ![](images/hotness_adjusted.png){width=45%}

</center>

I use a variety techniques to make adjustments to the Geek rating on boardgamegeek.First, I compute **Complexity-Adjusted Ratings** in order to identify which games get the most bang for their complexity buck. I am not the first to develop the method for computing these ratings, which is a very simple statistical model, but I do use that method to keep a list of up-to-date complexity ratings based on my latest pull of games from BGG.
Second, I amend the formula for the geek rating to place a greater weight on the popularity of the game, which I have found can mitigate a tendency of the geek ratings to skew towards the 'the hotness'.

## Reports

The following links display some quick analyses for the data that is available from boardgamegeek:

[**Community Ratings**](https://phenrickson.github.io/bgg_reports/reports/exploratory/outcomes.html)

[**Mechanics**](https://phenrickson.github.io/bgg_reports/reports/exploratory/mechanics.html)

[**Categories**](https://phenrickson.github.io/bgg_reports/reports/exploratory/categories.html)

[**Families**](https://phenrickson.github.io/bgg_reports/reports/exploratory/families.html)

[**Designers**](https://phenrickson.github.io/bgg_reports/reports/exploratory/designers.html)

[**Publishers**](https://phenrickson.github.io/bgg_reports/reports/exploratory/publishers.html)

---