shoni commited on
Commit
f89152e
·
1 Parent(s): ff20cb7

Update README with threads info

Browse files
Files changed (1) hide show
  1. README.md +81 -0
README.md CHANGED
@@ -1,3 +1,84 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
  license: cc-by-4.0
3
  ---
 
1
+ # Farcaster Threads Dataset
2
+
3
+ ## Overview
4
+
5
+ This dataset contains high-quality thread data from Farcaster's entire history, featuring 512-dimensional embeddings generated using Voyager (float32) on formatted thread text. The dataset includes comprehensive thread metadata, engagement metrics, and vector embeddings suitable for semantic search, recommendation systems, and content analysis.
6
+
7
+ ## Dataset Details
8
+
9
+ - **Total Records**: ~24.3 million threads
10
+ - **Data Cutoff**: 2025-05-21 19:44:58.000 -0600 (no threads newer than this date)
11
+ - **Quality Filter**: Non-spam content only (spam label = 2)
12
+ - **Embedding Model**: Voyager 512-dimensional float32 vectors
13
+ - **Repository**: [shoni/farcaster](https://huggingface.co/datasets/shoni/farcaster)
14
+ - **Data Location**: `/threads` folder
15
+
16
+ ## Schema
17
+
18
+ | Column | Type | Description |
19
+ |--------|------|-------------|
20
+ | `hash` | bytes | Unique thread identifier (primary key) |
21
+ | `fids` | list[int] | FIDs of reply authors, ordered by reaction count (most reacted first) |
22
+ | `reactions` | int | Total reaction count from non-spam users |
23
+ | `author_fid` | int | FID of the original thread author |
24
+ | `timestamp` | timestamp | Thread creation date |
25
+ | `claimed_at` | timestamp | When thread was claimed by processing worker |
26
+ | `blob` | string | Formatted thread text including replies and author information |
27
+ | `blob_embedding` | list[float32] | 512-dimensional Voyager embeddings of the blob text |
28
+
29
+ ## Data Quality
30
+
31
+ - **Spam Filtering**: Only includes threads with spam label = 2 (verified non-spam)
32
+ - **Engagement Metrics**: Reaction counts filtered to exclude spam users
33
+ - **Historical Coverage**: Complete Farcaster thread history up to cutoff date
34
+ - **Reply Ordering**: Reply authors (`fids`) sorted by engagement for relevance
35
+
36
+ ## Use Cases
37
+
38
+ - Semantic search over Farcaster content
39
+ - Thread recommendation systems
40
+ - Content similarity analysis
41
+ - Social network analysis
42
+ - Engagement prediction modeling
43
+ - Community detection and analysis
44
+
45
+ ## Loading the Dataset
46
+
47
+ ```python
48
+ from datasets import load_dataset
49
+
50
+ # Load the dataset
51
+ dataset = load_dataset("shoni/farcaster", data_dir="threads")
52
+
53
+ # Access embeddings
54
+ embeddings = dataset['train']['blob_embedding']
55
+
56
+ # Access formatted thread text
57
+ threads = dataset['train']['blob']
58
+ ```
59
+
60
+ ## Citation
61
+
62
+ If you use this dataset in your research, please cite:
63
+
64
+ ```
65
+ @dataset{farcaster_threads_2025,
66
+ title={Farcaster Threads Dataset with Voyager Embeddings},
67
+ author={shoni},
68
+ year={2025},
69
+ url={https://huggingface.co/datasets/shoni/farcaster}
70
+ }
71
+ ```
72
+
73
+ ## License
74
+
75
+ CC-BY-4.0
76
+
77
+ ## Repository
78
+
79
+ [https://huggingface.co/datasets/shoni/farcaster](https://huggingface.co/datasets/shoni/farcaster)
80
+
81
+
82
  ---
83
  license: cc-by-4.0
84
  ---