Hey Bryan, Thanks for this - will test and update the code and site soon.
Would love to know more about your project and methodology. It's clear from your messages and your Github account that this is some Human / AI collaboration experiment. To be clear, I appreciate your messages and fixes, but in the spirit of community and transparency, would really appreciate a quick introduction and what you're trying to do here - are you trawling multiple open data lists and having AI scan for issues and help out? Is it just datameet? Is this part of some organized effort? Some background would really help - Datameet is also a community of human beings, and it's helpful to have some background and some context of the human being(s) behind AI-assisted work. If you plan on contributing more regularly to the list, it would be really great if you could start a new thread with a brief introduction of your work and methodology and how you're using AI to make contributions, etc. Ofc, this is not a requirement, and your work is appreciated - I just feel in this world of increasingly AI-generated text, it's really nice to have a sense of the human intention behind some of this - otherwise we will end up with mailing lists with agents speaking to agents (which maybe ok?). Thanks again for your message and contribution - if you are comfortable sharing, would really appreciate a bit of background around the work you're doing. Cheers, Sanjay On Wed, Sep 2, 2026 at 8:50 AM Bryan <[email protected]> wrote: > Hi Rahul and Sanjay, > > I worked through the reconstruction and prepared a concrete correction, > with the before/after file and reproducible tests: > https://github.com/batpad/goamines/issues/1 > > Two date boundaries affect the result: movements on 31 March 2024 were > being added to that day's closing-stock baseline, while timestamped > movements on 31 March 2025 were excluded by the upper-bound comparison. The > patch starts after the opening closing-stock day and includes the whole > final day. > > On the downloaded public database, four reconstructed balances change. All > 166 patched reconstructions match an independent calendar-date calculation. > The issue contains the four-row report and a free download with the patch, > ten boundary tests, full receipt and instructions. The test runs in memory > and leaves the source database unchanged. > > This uses an end-of-day closing-stock interpretation. It does not resolve > missing baselines or source completeness, and negative values alone do not > establish physical shortage or wrongdoing. The royalty stream still has no > 2024 baseline. > > Thank you for making the data and code available — happy to discuss the > cutoff convention or adapt the patch. > Bryan / CivicDataForge > > -- > Datameet is a community of Data Science enthusiasts in India. Know more > about us by visiting http://datameet.org > --- > You received this message because you are subscribed to the Google Groups > "datameet" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > To view this discussion visit > https://groups.google.com/d/msgid/datameet/1d10abff-c260-47c2-b0ca-4051d7e61345n%40googlegroups.com > <https://groups.google.com/d/msgid/datameet/1d10abff-c260-47c2-b0ca-4051d7e61345n%40googlegroups.com?utm_medium=email&utm_source=footer> > . > -- Datameet is a community of Data Science enthusiasts in India. Know more about us by visiting http://datameet.org --- You received this message because you are subscribed to the Google Groups "datameet" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion visit https://groups.google.com/d/msgid/datameet/CAG3W7ZEpBDScGd85F4vEctWfV06_zyDf0_TfsrkYKV6Fdcv8nQ%40mail.gmail.com.
