otl-year-built

Leaflet map of parcels by decade built to display history of property development in West Hartford CT and Hamden CT, for https://OnTheLine.trincoll.edu

West Hartford map

https://ontheline.github.io/otl-year-built/index-caption.html

Hamden map

https://ontheline.github.io/otl-year-built/hamden/index-caption.html

Credits

Interactive map developed by Jack Dougherty and Ilya Ilyankou / Picturedigits for On The Line, http://OnTheLine.trincoll.edu, Trinity College, Hartford CT

West Hartford Data and Notes

Shapefiles with parcel identifiers and year-built data provided by Town of West Hartford and AppGeo in June 2021 from the Online Assessment Database https://gis.vgsi.com/westhartfordct/. Thanks to Jared Morin (jared.morin@westhartfordct.gov) and Rebecca Talamini (spatialiqsupport@appgeo.com).

Note that "yearbuilt" means "most recent year of construction recorded for this parcel" because this present-day assessment database was not designed to be a complete historical record of all property development. For example, it does not record when a building was first erected on a property (e.g. 1910), then torn down (e.g. 1970), because only the most recent date of construction is stored (e.g. 1970). Also, "yearbuilt" data missing for about 600 out of nearly 20,000 parcels (3%), which are included in a separate file no-year-built.geojson but not displayed in the interactive map. Only properties that currently exist are shown in this present-day assessment database.

Based on "yearbuilt", we added a new field named "decadebuilt" where "1900s" means 1900-1909, "1910s" means 1910-1919, etc. "Pre1900" means all years before 1900.

GIS_Area is measured in acres (eg. 4 Frederick Road = 0.16 acres)

West Hartford Data Cleaning and GeoJSON Conversion

Shapefiles processed by Jack Dougherty in https://Mapshaper.org to clean duplicates (condos and other units on same parcel with same or similar street address), create decadebuilt column, and to convert to GeoJSON files.

• Uploaded shapefiles.zip to Mapshaper.org
• ignore 23 line intersections
• console: -proj wgs84
• console: -split ParcelType, which divided into different files by type:
  • 20,042 Parcel
  • 2,387 Condo Main
  • 22 unnamed
  • 1 Water (ignore)
  • 1 Private ROW (ignore)
  • 1 Paper Street (ignore)
• Export into separate geojson files
• Also, export Condo Main into CSV and create pivot table by ParcelID and yearbuilt. Note that 56 features have multiple entries (units on same lot) but same ParcelID and same GIS_Area (where each condo is listed as the entire property lot, so needed to eliminate duplicates)
• Upload only the Condo predissolved geojson file, where multiple condos have same ParcelID (but slightly different addresses), and use console to dissolve but "copy" (aka keep) first instance of selected fields into the dissolved feature.
  • console: $ dissolve ParcelID copy-fields=ParcelID,ParcelType,siteaddres,GIS_Area,yearbuilt
  • [dissolve] Dissolved 2,387 features into 56 features
• Manually assign decadebuilt field to Condo CSV, and join to Condo GeoJSON
• Upload only Parcels predissolved geojson, where a few hundred multiple units have same ParcelID (similar to condos), and dissolve but keep first instance as above.
  • $ dissolve ParcelID copy-fields=ParcelID,ParcelType,siteaddres,GIS_Area,yearbuilt
  • [dissolve] Dissolved 20,042 features into 19,865 features
• Upload only unnamed GeoJSON, export CSV, and join in decadebuilt field. About 20 rows are empty and must be ignored.
• -merge-layers target=WH-Parcel-decadebuilt,WH-Condo-decadebuilt,WH-unnamed-decadebuilt
• simplified 100MB fullsize GeoJSON down 50% (large, 50MB) in order to upload to GitHub, and simplified at 0.5% (simplified, 11MB) 11MB, which you think would be even smaller, but it's not
• -split decadebuilt (to produce one GeoJSON file for each decade/century)
• downloaded WH-yearbuilt.csv and created pivot table to show results

West Hartford Results

Decade-Century Built	Count	Percent
1700	25	0%
1800	25	0%
1850	162	1%
1900	205	1%
1910	784	4%
1920	3019	15%
1930	2183	11%
1940	3020	15%
1950	6199	31%
1960	2181	11%
1970	742	4%
1980	306	2%
1990	179	1%
2000	206	1%
2010	90	0%
2020	7	0%
na	594	3%
Total	19927	100%

Decade-Century Built	Acres	Percent
1700	24	0%
1800	117	1%
1850	78	1%
1900	111	1%
1910	248	2%
1920	797	6%
1930	810	6%
1940	940	8%
1950	2949	24%
1960	3123	25%
1970	560	4%
1980	309	2%
1990	172	1%
2000	407	3%
2010	145	1%
2020	3	0%
na	1721	14%
Total	12517	100%

Acres based on sum of GIS_Area

Hamden data and cleaning

• Data provided by Andrew Kinlock, Town of Hamden GIS coordinator, Feb 2024
• created decadebuilt column in Excel
• removed about 20 rows from SRCOG_LU
  • Open Space
  • Other (golf course, cemetery, etc)
• confirmed UniqueID for each parcel, no duplicates, so simpler than WH data
• upload original hamden-parcels.geojson to Mapshaper.org, ignore Check Line Intersections, but BEFORE IMPORT in the box declare string-fields=id (to convert numerical id to string id to match column type inside geojson)
• join data keys=id,id
  • [join] Joined data from 15,605 source records to 15,605 target records
  • [join] 1149 target records received no data
• exported copy of hamden-parcels-largejoin.geojson as backup, then simplified 13MB file down to 21 percent, now 9MB (you think it would be smaller, but no)
• split decadebuilt (to produce one GeoJSON file for each decade) and exported separate geojson files, except the "null" file with no decadebuilt data. Largest file was 1950s at 1.9MB
• note how named Hamden decadebuilt and filenames differently (plural pre1900s) than WestHartford (singular pre1900)

Hamden results

Hamden Parcels (where yearbuilt does not equal zero)

DecadeBuilt	Grand Total	Residential	Pct Res Tot
pre1900s	259	224	2%
1900s	216	177	1%
1910s	288	266	2%
1920s	2215	2090	14%
1930s	1468	1403	10%
1940s	2369	2297	16%
1950s	3662	3522	24%
1960s	1856	1739	12%
1970s	1034	957	6%
1980s	1068	997	7%
1990s	670	624	4%
2000s	457	406	3%
2010s	64	39	0%
Grand Total	15626	14741	100%