U.S. Granted Utility Patents Originating in Saginaw-Saginaw Township North, MI (MSA) (PATENTCBSA140980) | FRED | St. Louis Fed
- Rows
- 16
- Columns
- 2
- Data range
- 2000-01-01 to 2015-01-01
Graph and download economic data for U.S. Granted Utility Patents Originating in Saginaw-Saginaw Township North, MI (MSA) (PATENTCBSA140980) from 2000 to 2015 about Saginaw, patent granted, patents, intellectual property, origination, MI, and USA.
AI analysis
Dataset Analysis: U.S. Granted Utility Patents — Saginaw-Saginaw Township North, MI (MSA)
1. What This Dataset Represents
This dataset tracks annual counts of granted utility patents originating from the Saginaw-Saginaw Township North, MI metropolitan statistical area (MSA), spanning 2000–2015 (16 annual observations). Sourced from the Federal Reserve Economic Data (FRED) API via the St. Louis Fed (fredgraph.csv?id=PATENTCBSA140980), this is a curated, publicly trusted economic indicator series. FRED's role as an aggregator means the underlying patent data likely originates from the USPTO, repackaged for economic research purposes. The dataset's value lies in linking innovation activity (a leading indicator of regional economic health) to other MSA-level economic series—unemployment, GDP, manufacturing output, or R&D investment—for a small industrial Michigan metro area. Given the annual granularity and 15-year span, this is best suited for long-run trend and regional comparison studies rather than high-frequency correlation work.
2. Data Quality Observations
The dataset is exceptionally clean at face value: zero null cells across both columns, and the date join key shows 16 distinct values matching 16 total rows, indicating no duplicate timestamps. However, two caveats temper confidence: (a) the duplicate-row check is marked "pending Phase D recomputation," so uniqueness at the full-row level hasn't been formally verified against the Parquet backend; (b) the sample data section is empty ("rows pending Phase D R2/Parquet wire-up"), meaning we cannot visually inspect actual values to sanity-check the summary statistics. For a dataset this small (16 rows), even minor data entry issues could meaningfully skew correlation results, so I'd recommend validating the raw CSV directly against FRED before downstream use.
3. Key Column Distributions
The patent count column (PATENTCBSA140980) ranges from 26 to 64 patents/year, with a mean of 47.4 and median of 52—the median exceeding the mean suggests a left-skewed distribution, confirmed by the negative skewness coefficient (-0.289), indicating more years with above-average counts and a few lower-count years pulling the mean down (likely early-2000s or recession-era troughs). The interquartile range (Q1=34.75, Q3=58.5) spans about 23.75 patents, and with only 13 distinct values across 16 rows, there are a couple of repeated counts—worth checking whether these repeats are coincidental or data artifacts. The standard deviation of 13.47 relative to the mean (~28% coefficient of variation) indicates moderate year-to-year volatility, consistent with a small-MSA patent series where a handful of firms' filing activity can swing annual totals.
4. Recommended Join Key
The observationdate column (annual, January 1 of each year) is the natural and only viable join key. Since it's already a clean Date type with zero nulls and full distinctness, it should join reliably against any other FRED series using the same annual date convention. Analysts should ensure date alignment (e.g., fiscal year vs. calendar year) when joining against non-FRED sources that may report mid-year or quarterly dates requiring resampling to annual.
5. Suggested Companion Datasets
For meaningful correlation analysis, pair this with: (1) Saginaw MSA unemployment rate or labor force data (FRED series like SAGI... UR) to test whether patent activity leads/lags employment cycles; (2) regional GDP or manufacturing output given Saginaw's automotive/manufacturing heritage; (3) R&D expenditure or venture capital data at the state or metro level; (4) national/state-level utility patent trends to distinguish local innovation shifts from broader USPTO filing trends; and (5) population or establishment counts (Census/BLS QCEW) to normalize patents per capita or per firm, since raw counts in a small MSA are highly sensitive to a few large employers' R&D activity.
Columns
- observation_date (date)
- PATENTCBSA140980 (integer)