thomaswmitch commited on
Commit
2ce5d0a
·
verified ·
1 Parent(s): 580968e

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +57 -29
README.md CHANGED
@@ -1,31 +1,59 @@
1
  ---
2
- dataset_info:
3
- features:
4
- - name: trade_id
5
- dtype: string
6
- - name: ticker
7
- dtype: string
8
- - name: count
9
- dtype: int64
10
- - name: created_time
11
- dtype: string
12
- - name: yes_price
13
- dtype: int64
14
- - name: no_price
15
- dtype: int64
16
- - name: taker_side
17
- dtype: string
18
- - name: market_ticker
19
- dtype: string
20
- splits:
21
- - name: train
22
- num_bytes: 817847758
23
- num_examples: 5076511
24
- download_size: 276868855
25
- dataset_size: 817847758
26
- configs:
27
- - config_name: default
28
- data_files:
29
- - split: train
30
- path: data/train-*
31
  ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
+ license: mit
3
+ pretty_name: Kalshi Trades (Aug 2023–Aug 2025)
4
+ tags:
5
+ - prediction-markets
6
+ - kalshi
7
+ - trades
8
+ - economics
9
+ - finance
10
+ - time-series
11
+ language:
12
+ - en
13
+ task_categories:
14
+ - other
15
+ size_categories:
16
+ - 1M<n<10M
 
 
 
 
 
 
 
 
 
 
 
 
 
 
17
  ---
18
+
19
+ # Kalshi Prediction Markets — Trades
20
+
21
+ High-volume trade-level data from Kalshi prediction markets spanning **Aug 2023–Aug 2025**, suitable for market microstructure, liquidity, and price-impact analysis.
22
+
23
+ > **Rows:** ~5.08M trades (single split)
24
+ > **Schema stability:** stable
25
+ > **Privacy:** public market data, no PII
26
+
27
+ ---
28
+
29
+ ## Dataset Structure
30
+
31
+ ### Split
32
+ - **train** — all trades
33
+
34
+ ### Features (columns)
35
+
36
+ | name | dtype | description |
37
+ |----------------|---------|-----------------------------------------------------------------------------|
38
+ | `trade_id` | string | Unique trade identifier |
39
+ | `ticker` | string | Contract/series symbol at the trade level |
40
+ | `market_ticker`| string | Parent market symbol |
41
+ | `count` | int64 | Matched quantity (units/contracts) |
42
+ | `created_time` | string | UTC timestamp of trade (ISO-8601 string) |
43
+ | `yes_price` | int64 | Price for **YES** leg (integer; typically cents or ticks) |
44
+ | `no_price` | int64 | Price for **NO** leg (integer; typically cents or ticks) |
45
+ | `taker_side` | string | Aggressor side for the trade (string enum; e.g., `"buy"`/`"sell"` or similar)|
46
+
47
+ > **Note:** `yes_price`/`no_price` are stored as integers. If your downstream pipeline expects dollars, convert as needed (e.g., divide by `100` if prices are in cents).
48
+
49
+ ---
50
+
51
+ ## Usage
52
+
53
+ ### Load with `datasets`
54
+ ```python
55
+ from datasets import load_dataset
56
+
57
+ ds = load_dataset("thomaswmitch/kalshi-prediction-markets-trades") # update to your repo name if different
58
+ print(ds)
59
+ print(ds["train"].features)