The data, in full

Name: Dholera Progress Record, full dataset
Creator: Bhavik Sarkhedi
License: https://creativecommons.org/licenses/by/4.0/

Both datasets behind this site are downloadable in full, 115 monthly records and 225 annual report records, as JSON and CSV, licensed Creative Commons Attribution 4.0 International. The gap lists and the verification metadata are included, not stripped out.

The whole register, openly licensed

Downloads
File	Format	Extent	What it contains
dmu-progress-timeseries.json	JSON	115 records	The monthly series exactly as extracted, including the gap list and the qualitative record.
dicdl-annual-report.json	JSON	225 records	The audited annual report dataset, including its verification block, gaps and caveats.
metric-series.csv	CSV	340 rows	Both datasets in one long table: dataset, metric, date, value, unit, tag, source, page.
dmu-readings.csv	CSV	115 rows	The monthly series as a flat table with every field.
dicdl-facts.csv	CSV	225 rows	The annual report dataset as a flat table with every field.

The compilation is licensed Creative Commons Attribution 4.0 International, so it can be reused with attribution to this register. The underlying documents are publications of the Government of India and the Government of Gujarat and carry their own terms. Nothing here is behind a form, an email gate or a login.

Licence: Creative Commons Attribution 4.0 International. Attribution: Dholera Progress Record, https://dholera.today/. Compiled by Bhavik Sarkhedi.

Field list, monthly series

Every record in the monthly file carries these 15 fields. Two of them exist so that a reader can see the difference between the value as a number and the value as the document printed it.

metric
metric_verbatim_in_source
value
value_as_printed
unit
scope
as_of_date
as_of_date_as_printed
source_file
source_page
source_section
tag
note
value_min
value_max

Source: field list declared in the monthly dataset file.

Field list, annual reports

The company file carries 28 fields, because a contract record needs more of them than a financial line does. Fields that do not apply to a record are null rather than blank.

fact_name
value
unit
financial_year
date_reference
source_file
source_page
source_page_printed
tag
extraction
verbatim
note
designation
din
shares
dates
contractor
contract_basis
award_date
project_cost_inr_crore
project_cost_note
management_fee_inr_crore
project_management_consultant
plot_area_sq_m
built_up_area_sq_m
construction_start_date
specification
status

Source: field list declared in the company dataset file.

What you should not do with it

Do not present a TARGET row as an outcome. Do not average two readings that disagree. Do not fill a missing month by interpolation and then attribute the result to this register. Every one of those moves would produce a number that no government document has ever published, which is the failure mode this site exists to avoid.

Editorial rule of this register.