dylanalloy commited on
Commit
759b2fd
1 Parent(s): 445982d
Files changed (1) hide show
  1. README.md +94 -0
README.md CHANGED
@@ -1,3 +1,97 @@
1
  ---
2
  license: cc-by-nc-4.0
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
3
  ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
  license: cc-by-nc-4.0
3
+ task_categories:
4
+ - text-generation
5
+ language:
6
+ - en
7
+ tags:
8
+ - finance
9
+ - legal
10
+ pretty_name: swan - finance dataset
11
+ size_categories:
12
+ - 100K<n<1M
13
+ configs:
14
+ - config_name: default
15
+ data_files:
16
+ - split: corpus
17
+ path: "corpus.txt"
18
+ - split: corpus_deduped
19
+ path: "corpus_deduped.txt"
20
+ - split: sec_tracker
21
+ path: "all_sec_filings.csv"
22
+ - split: leaked_tracker
23
+ path: "all_leaked_pdfs.csv"
24
+ - split: fed_tracker
25
+ path: "all_fed_filings.csv"
26
+ - split: bls_jolts_tracker
27
+ path: "all_bls_jolts.csv"
28
+ - split: bls_cpi_tracker
29
+ path: "all_bls_cpi.csv"
30
+ - split: bls_ces_tracker
31
+ path: "all_bls_ces.csv"
32
+ - split: bls_historical_tracker
33
+ path: "all_bls_historical.csv"
34
  ---
35
+ <!-- header start -->
36
+ <div style="min-width:100%">
37
+ <center>
38
+ <img style="max-width:200px" src="https://huggingface.co/datasets/dylanalloy/swan/resolve/main/swan.png">
39
+ <h3>swan</h3>
40
+ <small>aggressively updated financial text dataset</small>
41
+
42
+ <a href="https://github.com/DylanAlloy/swan_scrape" target="_blank">scraping code</a>
43
+
44
+ </center>
45
+ </div>
46
+ <!-- header end -->
47
+
48
+
49
+ ### usage
50
+
51
+ ```python
52
+ from datasets import load_dataset
53
+ sets = ["corpus", "corpus_deduped"]
54
+ swan_data, swan_deduped = [load_dataset("dylanalloy/swan", data_files=f"{_}.txt") for _ in sets]
55
+ swan_data, swan_deduped
56
+ ```
57
+
58
+ ### data
59
+
60
+ <center>
61
+
62
+ | data | added |
63
+ | ----------- | ----------- |
64
+ | SEC filings | Wed. Aug 30th, 2023 |
65
+ | Federal Reserve transcripts | Wed. Aug 30th, 2023 |
66
+ | private wealth management releases | Wed. Aug 30th, 2023 |
67
+ | large bank releases | Wed. Aug 30th, 2023 |
68
+ | large fund releases | Wed. Aug 30th, 2023 |
69
+ | large trading firm releases | Wed. Aug 30th, 2023 |
70
+ | BLS JOLTS releases | Wed. Aug 30th, 2023 |
71
+ | BLS CPI releases | Wed. Aug 30th, 2023 |
72
+ | BLS CES releases | Wed. Aug 30th, 2023 |
73
+ | BLS historical reports | Wed. Aug 30th, 2023 |
74
+ </center>
75
+
76
+ ### updates
77
+
78
+ <small>this repo updates daily at 6AM EST</small>
79
+
80
+ | SEC Filngs | Federal Reserve transcripts | releases & reports |
81
+ | :--- | :----: | ---: |
82
+ | 30 minutes | daily | daily |
83
+
84
+ <small>🐒 **corpus** ⌨️ updated daily</small>
85
+
86
+ ### stats and delta
87
+
88
+ <center>
89
+ <img style="max-width:100%" src="https://huggingface.co/datasets/dylanalloy/swan/resolve/main/words_sizes.png">
90
+ <img style="max-width:100%" src="https://huggingface.co/datasets/dylanalloy/swan/resolve/main/vocab_sizes_time.png">
91
+ </center>
92
+
93
+ ### organization
94
+
95
+ - *.csv: tracker
96
+ - corpus.txt: collated text from all documents across all categories (designed for base model training)
97
+ - corpus_deduped.txt: unique lines of corpus