Datasets:
Update README.md
Browse files
README.md
CHANGED
|
@@ -4,10 +4,13 @@ language:
|
|
| 4 |
pretty_name: GeoComp
|
| 5 |
tags:
|
| 6 |
- GeoLocation
|
|
|
|
| 7 |
---
|
| 8 |
|
| 9 |
# GeoComp
|
|
|
|
| 10 |
## Dataset description
|
|
|
|
| 11 |
Inspired by [geoguessr.com](https://www.geoguessr.com/), we developed a free geolocation game platform that tracks participants' competition histories.
|
| 12 |
Unlike most geolocation websites, including Geoguessr, which rely solely on samples from Google Street View, our platform integrates Baidu Maps and Gaode Maps to address coverage gaps in regions like mainland China, ensuring broader global accessibility.
|
| 13 |
The platform offers various engaging competition modes to enhance user experience, such as team contests and solo matches.
|
|
@@ -20,34 +23,47 @@ To prevent cheating, external search engines are banned, and each round is time-
|
|
| 20 |
To ensure predictions are human-generated rather than machine-generated, users must register with a phone number, enabling tracking of individual activities.
|
| 21 |
Using this platform, we collected **GeoComp**, a comprehensive dataset covering 1,000 days of user competition.
|
| 22 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 23 |
## Requirement
|
|
|
|
| 24 |
The **GeoComp** is only for reasearch.
|
| 25 |
|
| 26 |
## Start
|
| 27 |
|
| 28 |
-
### Get csv
|
|
|
|
| 29 |
Merge the splited files to tuxun_combined.csv
|
|
|
|
| 30 |
```shell
|
| 31 |
cat tuxun_comblined_* > tuxun_comblined.csv
|
| 32 |
|
| 33 |
ls -lh tuxun_comblined.csv
|
| 34 |
```
|
| 35 |
|
| 36 |
-
### Data format
|
| 37 |
|
| 38 |
#### Example
|
| 39 |
-
|
| 40 |
-
|
| 41 |
-
|
|
|
|
|
| 42 |
|
| 43 |
#### Explanation
|
| 44 |
-
We hide data items that may reveal personal privacy like changing the value of key "userId" to "User", "hostUserId" to "HostUser", "playerIds" to "Players", "id" to "Game"
|
| 45 |
|
| 46 |
-
|
|
|
|
|
|
|
| 47 |
|
| 48 |
## Additional Information
|
| 49 |
|
| 50 |
### Citation Information
|
|
|
|
| 51 |
```bibtex
|
| 52 |
@misc{song2025geolocationrealhumangameplay,
|
| 53 |
title={Geolocation with Real Human Gameplay Data: A Large-Scale Dataset and Human-Like Reasoning Framework},
|
|
@@ -61,6 +77,7 @@ The data under the "data" column is in json style
|
|
| 61 |
```
|
| 62 |
|
| 63 |
### Links
|
|
|
|
| 64 |
[arXiv](https://arxiv.org/abs/2502.13759)
|
| 65 |
|
| 66 |
[Hugging Face](https://huggingface.co/papers/2502.13759)
|
|
|
|
| 4 |
pretty_name: GeoComp
|
| 5 |
tags:
|
| 6 |
- GeoLocation
|
| 7 |
+
|
| 8 |
---
|
| 9 |
|
| 10 |
# GeoComp
|
| 11 |
+
|
| 12 |
## Dataset description
|
| 13 |
+
|
| 14 |
Inspired by [geoguessr.com](https://www.geoguessr.com/), we developed a free geolocation game platform that tracks participants' competition histories.
|
| 15 |
Unlike most geolocation websites, including Geoguessr, which rely solely on samples from Google Street View, our platform integrates Baidu Maps and Gaode Maps to address coverage gaps in regions like mainland China, ensuring broader global accessibility.
|
| 16 |
The platform offers various engaging competition modes to enhance user experience, such as team contests and solo matches.
|
|
|
|
| 23 |
To ensure predictions are human-generated rather than machine-generated, users must register with a phone number, enabling tracking of individual activities.
|
| 24 |
Using this platform, we collected **GeoComp**, a comprehensive dataset covering 1,000 days of user competition.
|
| 25 |
|
| 26 |
+
## File introduction
|
| 27 |
+
|
| 28 |
+
- [tuxun_combined_*](https://huggingface.co/datasets/ShirohAO/tuxun/blob/main/tuxun_combined_aa): The splited files of tuxun_combined.csv, you can use "cat" to get the csv file.
|
| 29 |
+
- [tuxun_sample.csv](https://huggingface.co/datasets/ShirohAO/tuxun/blob/main/tuxun_sample.csv): An example to preview the structure of tuxun_combined.csv.
|
| 30 |
+
- [selected_panoids.csv](https://huggingface.co/datasets/ShirohAO/tuxun/blob/main/selected_panoids.csv): The 500 panoids we used in our work.
|
| 31 |
+
- [download_panoramas.py](https://huggingface.co/datasets/ShirohAO/tuxun/blob/main/download_panoramas.py): The script to download street view images from the panoid.
|
| 32 |
+
|
| 33 |
## Requirement
|
| 34 |
+
|
| 35 |
The **GeoComp** is only for reasearch.
|
| 36 |
|
| 37 |
## Start
|
| 38 |
|
| 39 |
+
### Get tuxun_combined.csv
|
| 40 |
+
|
| 41 |
Merge the splited files to tuxun_combined.csv
|
| 42 |
+
|
| 43 |
```shell
|
| 44 |
cat tuxun_comblined_* > tuxun_comblined.csv
|
| 45 |
|
| 46 |
ls -lh tuxun_comblined.csv
|
| 47 |
```
|
| 48 |
|
| 49 |
+
### Data format of tuxun_combined.csv
|
| 50 |
|
| 51 |
#### Example
|
| 52 |
+
|
| 53 |
+
| id | data | gmt_create | timestamp |
|
| 54 |
+
| ---- | ------------------------------------------------------------ | --------------- | --------- |
|
| 55 |
+
| Game | "{""id"": ""Game"", ""type"": ""map_country_streak"", ""gmtCreate"": 1734188074762, ""gmtFinish"": 1734188097294, ""hostUserId"": ""HostUser"", ""playerIds"": ""Players"", ""numPlayersJoined"": 0, ""status"": ""finish"", ""currentRoundNumber"": 1, ""roundNumber"": 0, ""rounds"": [{""round"": 1, ""id"": ""Game"", ""contentType"": ""panorama"", ""lat"": -27.1405249973, ""lng"": -59.4525319805, ""heading"": 290.8791503906, ""startTime"": 1734188075183, ""endTime"": 1734188097294, ""isDamageMultiple"": false, ""damageMultiple"": 1.0, ""nation"": ""阿根廷"", ""move"": false, ""source"": ""google_pano"", ""panoId"": ""ZF8FdQbXVt4XYSgP9szamg"", ""vHeading"": 290.74, ""vZoom"": 0.0, ""vPitch"": 0.07, ""pan"": true, ""zoom"": true}], ""teams"": [], ""teamsSize"": 2, ""player"": {""streaks"": 0, ""lastRoundResult"": {""round"": 1, ""score"": 15, ""distance"": 11616.42890082367, ""guessPlace"": ""挪威"", ""targetPlace"": ""阿根廷""}, ""roundResults"": [], ""totalScore"": 15, ""userId"": ""User"", ""guesses"": [{""round"": 1, ""gmtCreate"": 1734188096396, ""lat"": 61.52588100795455, ""lng"": 9.678774290625825, ""distance"": 11616.42890082367, ""timeConsume"": 21213, ""score"": 15, ""type"": ""guess""}], ""pins"": [{""round"": 1, ""gmtCreate"": 1734188095256, ""lat"": 60.68033874641395, ""lng"": 14.111789398216388, ""timeConsume"": 20073, ""type"": ""guess""}, {""round"": 1, ""gmtCreate"": 1734188096124, ""lat"": 61.52588100795455, ""lng"": 9.678774290625825, ""timeConsume"": 20941, ""type"": ""guess""}]}, ""startTime"": 1734188074762, ""createTime"": 1734188074762, ""multiplierOpen"": true, ""streaks"": 0, ""leftSkipTimes"": 3, ""saveTeamCount"": 0, ""move"": false, ""moveType"": ""noMove"", ""mapsId"": 1418, ""mapsName"": ""西湖十景-苏堤春晓"", ""centerLng"": 9.3914097615, ""centerLat"": 51.2758208168, ""mapZoom"": 4, ""mapMaxLat"": 81.6812215034, ""mapMinLat"": -85.0000164841, ""mapMaxLng"": 178.3897528865, ""mapMinLng"": -177.3755914392, ""scoreDistance"": 1853.693498982, ""health"": 6000, ""pan"": true, ""zoom"": true, ""china"": false}" | 1734188074762.0 | |
|
| 56 |
|
| 57 |
#### Explanation
|
|
|
|
| 58 |
|
| 59 |
+
- We hide data items that may reveal personal privacy like changing the value of key "userId" to "User", "hostUserId" to "HostUser", "playerIds" to "Players", "id" to "Game"
|
| 60 |
+
|
| 61 |
+
- The data under the "data" column is in json style. This column contains the detailed geolocation information like "lat", "lng", "nation" and "panoId".
|
| 62 |
|
| 63 |
## Additional Information
|
| 64 |
|
| 65 |
### Citation Information
|
| 66 |
+
|
| 67 |
```bibtex
|
| 68 |
@misc{song2025geolocationrealhumangameplay,
|
| 69 |
title={Geolocation with Real Human Gameplay Data: A Large-Scale Dataset and Human-Like Reasoning Framework},
|
|
|
|
| 77 |
```
|
| 78 |
|
| 79 |
### Links
|
| 80 |
+
|
| 81 |
[arXiv](https://arxiv.org/abs/2502.13759)
|
| 82 |
|
| 83 |
[Hugging Face](https://huggingface.co/papers/2502.13759)
|