File size: 9,230 Bytes
74c3da6
 
 
a6654f5
 
74c3da6
 
 
 
 
a6654f5
 
74c3da6
 
 
 
a6654f5
74c3da6
 
a6654f5
74c3da6
a6654f5
74c3da6
a6654f5
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
df28a20
a6654f5
df28a20
a6654f5
df28a20
a6654f5
 
 
 
 
 
 
 
df28a20
 
 
a6654f5
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
df28a20
 
 
a6654f5
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
df28a20
 
a6654f5
 
 
 
 
 
 
df28a20
a6654f5
df28a20
a6654f5
df28a20
a6654f5
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
df28a20
 
a6654f5
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
---
pretty_name: "FinChrona: Open Financial Time Series"
license: cc0-1.0
language:
  - en
tags:
  - finance
  - economics
  - tabular
  - timeseries
  - treasury
  - interest-rates
configs:
  - config_name: default
    data_files:
      - split: train
        path: "finchrona_treasury_rates.parquet"
---

# FinChrona: Open Financial Time Series

*Open financial data, built for research.* **Release: v1.0.0** · **Data cutoff: 2026-10-23**

FinChrona is a curated historical dataset of daily U.S. Treasury nominal par yields, real/TIPS par yields, and Treasury bill rates, standardized into one consistent long-format schema.

- **Rows:** 180,815
- **Columns:** 7
- **Nominal par yields:** 1990-01-02 to 2026-10-23
- **Real/TIPS par yields:** observations begin in 2003
- **Treasury bill rates:** observations begin in 2002
- **Source retrieval date:** 2026-10-25

Coverage varies by rate series and maturity. The release is a historical snapshot, not a real-time feed.

## Get the dataset

- **Zenodo archive — v1.0.0:** [10.5281/zenodo.23730813](https://doi.org/10.5281/zenodo.23730813)
- **Zenodo concept DOI — all versions:** [10.5281/zenodo.23730812](https://doi.org/10.5281/zenodo.23730812)
- **GitHub repository:** https://github.com/itzsam-lol/finchroma_01
- **Kaggle dataset:** https://www.kaggle.com/datasets/samuraiiii/finchrona-financial-time-series
- **Kaggle EDA notebook:** https://www.kaggle.com/code/samuraiiii/finchrona-treasury-rates-eda

This Hugging Face repository is a distribution and discovery location for FinChrona. The Zenodo version DOI identifies the archived v1.0.0 research artifact; use it when citing this exact release.

## Series coverage

| `curve_type` | Observation coverage | Rows in v1.0.0 |
|---|---|---:|
| `par_nominal` | 1990-01-02 to 2026-10-23 | 99,894 |
| `par_real` | 2003-01-02 to 2026-10-23 | 27,609 |
| `bill_rate` | 2002-01-02 to 2026-10-23 | 53,312 |
| **Total** | Varies by series and maturity | **180,815** |

Bill-rate observations include `bank_discount` and `coupon_equivalent` quote conventions.

## Files

- `finchrona_treasury_rates.parquet` — primary analytical file.
- `finchrona_treasury_rates.csv` — full CSV version of the same observations.
- `sample_last_60_days.csv` — small sample for quick inspection.
- `data_dictionary.md` — column definitions, data types, units, and conventions.
- `methodology.md` — source acquisition and transformation procedure.
- `quality_report.md` — validation and coverage summary.
- `licensing_and_attribution.md` — licensing rationale and attribution notes.
- `release_manifest.json` — release metadata, file sizes, and SHA-256 checksums.

## Schema

Each row represents one rate observation.

| Column | Meaning |
|---|---|
| `observation_date` | U.S. Treasury business date associated with the observation (`YYYY-MM-DD`); not a publication timestamp. |
| `curve_type` | `par_nominal`, `par_real`, or `bill_rate`. |
| `quote_basis` | `par_yield`, `bank_discount`, or `coupon_equivalent`, as applicable. |
| `maturity_label` | Compact maturity identifier such as `4W`, `1M`, `1.5M`, `2Y`, or `10Y`. |
| `maturity_value` | Numeric maturity quantity, interpreted together with `maturity_unit`. |
| `maturity_unit` | `week`, `month`, or `year`. |
| `value_pct` | Annualized rate in percentage units; `3.74` means `3.74%`, not `0.0374`. |

**Primary key:** `(observation_date, curve_type, quote_basis, maturity_label)`.

See [`data_dictionary.md`](data_dictionary.md) for the complete data dictionary.

## Quickstart

Install pandas and a Parquet reader such as PyArrow, then load the Parquet file:

```python
import pandas as pd

path = "finchrona_treasury_rates.parquet"
df = pd.read_parquet(path)

print(df.shape)
print(df.head())
print(df.groupby("curve_type")["observation_date"].agg(["min", "max", "count"]))

nominal = df[
    (df["curve_type"] == "par_nominal")
    & (df["quote_basis"] == "par_yield")
]
wide = nominal.pivot(
    index="observation_date",
    columns="maturity_label",
    values="value_pct",
)

spread_10y_2y = wide["10Y"] - wide["2Y"]
print(spread_10y_2y.describe())
```

This example assumes the file has already been downloaded from the Hub or the Zenodo release. See the [Hugging Face download documentation](https://huggingface.co/docs/hub/datasets-downloading) for supported download methods.

## Source and methodology

**Source:** U.S. Department of the Treasury, [Daily Treasury Rates](https://home.treasury.gov/policy-issues/financing-the-government/interest-rate-statistics).

Source files were retrieved on 2026-10-25, with a data cutoff of 2026-10-23. The pipeline collects the published CSV tables, records source and file-integrity information, converts the original tables into long format, standardizes maturity labels and quote conventions, and validates the resulting schema and record keys.

Blank source cells are omitted rather than imputed or forward-filled. Negative real yields and bill-rate observations are retained where present in the source.

More detail is available in [`methodology.md`](methodology.md) and [`quality_report.md`](quality_report.md).

## Data quality

The v1.0.0 validation report records:

- 180,815 rows and 7 columns.
- No null values in the processed output.
- No duplicate primary keys.
- 180,815 non-null source cells reconciled to 180,815 output rows.
- 5,032 negative values in total: 4,940 real-yield observations and 92 bill-rate observations.

These checks describe the processed snapshot and do not guarantee that every source observation is error-free. See [`quality_report.md`](quality_report.md) for the reported checks and coverage.

## Limitations and responsible use

- **Not point-in-time:** the release does not preserve complete publication timestamps or a full history of revisions. `observation_date` must not be treated as the earliest time information was available.
- **No imputation:** blank source cells are omitted, so maturities appear only when present in the published source.
- **Model-derived rates:** Treasury par yields are curve points, not guaranteed executable trading prices. Bill rates retain their published quotation conventions.
- **Historical snapshot:** this is daily historical data, not a real-time market feed.
- **Coverage differences:** series start dates and available maturities vary. The 30-year nominal series has a source gap from 2002-02-19 through 2006-02-08, as documented in the source data.
- **No investment advice:** the dataset is for research and analysis, not trading recommendations.

For forecasting or backtesting, ensure that inputs would have been available at the simulated prediction time and use evaluation methods appropriate for time-series data.

## Licensing and attribution

The repository code is licensed under the MIT License; see the [GitHub repository](https://github.com/itzsam-lol/finchroma_01) for the code licence. The compiled dataset is designated CC0 1.0 in the accompanying licensing documentation.

The licensing documentation records that no explicit dataset-licence text was found on the Treasury pages inspected and explains the project's rationale and remaining uncertainty. Users should review `licensing_and_attribution.md` and applicable source terms, particularly for uses outside the United States. The MIT code licence does not replace the separately documented dataset terms.

Attribution requested: **Source: U.S. Department of the Treasury, Daily Treasury Rates; compiled and organized by FinChrona.** FinChrona is independently maintained and is not affiliated with or endorsed by the U.S. Treasury.

## Citation

For the exact v1.0.0 release, cite the Zenodo archived dataset:

> Satyam. (2026). *FinChrona v1.0.0: Open Financial Time Series | U.S. Treasury Rates, 1990–2026* (Version 1.0.0) [Dataset]. Zenodo. https://doi.org/10.5281/zenodo.23730813

- **Version DOI:** https://doi.org/10.5281/zenodo.23730813
- **Concept DOI (all versions):** https://doi.org/10.5281/zenodo.23730812

BibTeX:

```bibtex
@misc{satyam2026finchrona,
  author    = {Satyam},
  title     = {FinChrona v1.0.0: Open Financial Time Series | U.S. Treasury Rates, 1990--2026},
  year      = {2026},
  publisher = {Zenodo},
  version   = {1.0.0},
  doi       = {10.5281/zenodo.23730813},
  url       = {https://doi.org/10.5281/zenodo.23730813},
  note      = {Dataset}
}
```

The Zenodo version DOI identifies this archived release. If a native Hugging Face DOI is generated later, document it separately as an identifier for the corresponding Hub revision rather than silently replacing the Zenodo citation.

## Updates

FinChrona v1.0.0 has a data cutoff of 2026-10-23. The intended update cadence is quarterly and manual; this is a target, not an automated process or a guarantee. Future releases should document source-coverage changes, any source revisions, transformation changes, updated quality reports, and new file checksums.

## Feedback

Please use the [GitHub issues page](https://github.com/itzsam-lol/finchroma_01/issues) to report data issues or suggest improvements. Include the affected file, observation date or record key, and reproducible details where possible. Do not include credentials or private information in issue reports.