We compared all 35 tickets on the Cohorts Engineering board with what is actually working in the app today. In short: a researcher can find cohorts, look inside them, pick what data they want, see a price and put it in a cart. Everything after the cart is not built yet.
The app runs on our development environment only. On the live site, the Cohorts button still says "Coming soon".
| Ticket | What it's for | Status | Where it stands |
|---|---|---|---|
| ENG-35 | Sign up & sign in | Working | Handled by the existing Pan.bio account: sign-up, sign-in, email check and password reset already exist there. Missing: signing in through a university or hospital login, an institution field, and a sign-out button. |
| ENG-34 | Roles & permissions | Rethink | Nothing built. Our agreed plan changes this: the organization's owner and admins decide who can buy, instead of four separate roles. The ticket should be rewritten. |
| Ticket | What it's for | Status | Where it stands |
|---|---|---|---|
| ENG-33 | Browse the catalog | Partly | Grid and table views, sorting, paging through results. Cohort names and descriptions are generated from internal codes (for example "Tcga Brca Cohort"), there is no "last updated" date or open/restricted label, and "most recent" sorting is not real. |
| ENG-32 | Search | Working | Finds cohorts by disease, place, data type and more, and tolerates typos. It understands a short fixed list of medical abbreviations, not a full medical vocabulary. |
| ENG-31 | Filters | Partly | Ten filters, each showing how many cohorts match. There is no sample-size filter, and a "years of data" filter exists behind the scenes but isn't on screen. Several filtering bugs have been reported; this is the main focus of the QA round. |
| ENG-30 | Search in plain English | Not started | Typing a sentence like "large breast cancer cohorts in the Middle East" and having filters set for you. |
| Ticket | What it's for | Status | Where it stands |
|---|---|---|---|
| ENG-29 | Cohort page | Partly | Tabs with real numbers. Missing: who provides the data, open or restricted access, last updated, and a completeness score. The page can crash in some situations; the fix is written but not yet released. |
| ENG-28 | Charts & statistics | Partly | Diagnoses, age, sex, ancestry, data types, tissue types. Missing: survival curves, common co-existing conditions, downloading a chart. "Collection methods" is always empty, and tabs show even when a cohort has nothing for them. |
| ENG-6 | Only show what a cohort has | Not started | Hiding sections a dataset doesn't have. The layout is fixed today. Depends on ENG-5. |
| Ticket | What it's for | Status | Where it stands |
|---|---|---|---|
| ENG-27 | Build a data package | Working | Pick sample size, data types and how long access lasts, or start from a Discovery / Research / Enterprise preset. The presets live only in the screen, not on the server, so they can drift. |
| ENG-26 | Clear pricing | Partly | Itemised price with discounts for larger samples and longer access. Missing: a discount for bundling data types, a way for admins to change prices, "contact us" pricing and free open-access cohorts. The 12-month discount shown on screen (10%) doesn't match the price charged (30%). |
| ENG-25 | Cart | Working | Saved to the account, works across devices, items can be removed. Items can't be edited yet. The cart belongs to the person; our plan says it should belong to the organization. |
| Ticket | What it's for | Status | Where it stands |
|---|---|---|---|
| ENG-24 | Submit a data request | Not started | The "Request Access" button exists but is switched off. For the first release our plan replaces this with buy and pay. |
| ENG-23 | Review & approval | Postponed | Our plan skips a human review step in the first release; purchases go straight to a secure workspace. Comes back later. |
| ENG-22 | My requests | Not started | Becomes a "My purchases" page with a button to open the secure workspace. |
| ENG-21 | Secure environment | Not started | The secure research workspace where bought data is used and can't leave. Built on Yahia's track; Cohorts connects to it. |
| ENG-20 | Patient records browser | Not started | Inside the secure workspace. |
| ENG-19 | Genetic variant browser | Not started | Inside the secure workspace. |
| ENG-18 | Medical imaging browser | Not started | Inside the secure workspace. |
| Ticket | What it's for | Status | Where it stands |
|---|---|---|---|
| ENG-17 | Admin dashboard | Not started | Today only engineers can load data, using a shared key. |
| ENG-16 | Usage analytics | Not started | |
| ENG-15 | Data provider portal | Not started | |
| ENG-14 | Partner onboarding | Not started | Only a spreadsheet template for describing a dataset exists. |
| ENG-13 | Automatic data refresh | Partly | Engineers can pull public cancer data (TCGA) by hand. Nothing runs on a schedule, and there's no "last synced" date. |
| ENG-5 | List of fields per dataset | Not started | The record of which fields each dataset has, which drives ENG-6. |
| ENG-1 | KHCC metadata (100 patients) | Not started | No description on the ticket. Today only TCGA and synthetic test data can be loaded. |
| ENG-7–12 | Six placeholder ideas | Not started | Export, saved searches, comparing cohorts, request messages, email preferences, notification centre. None of these have a description yet. |
| Ticket | What it's for | Status | Where it stands |
|---|---|---|---|
| ENG-2, 3, 4 | Notebook machine follow-ups | Move | About notebook computers, not Cohorts. ENG-3 is done; ENG-2 and ENG-4 should move to the Notebooks project. |
Found while reading the code, before any hands-on testing. The QA round will add to this list.
Before building anything new, we test every working feature on the development site, with filtering as the main focus. Each problem goes on the board with steps to reproduce it and how serious it is.
Fix the bugs by severity, release the cohort-page crash fix, and correct the discount label. Rewrite the tickets our plan has changed (roles, requests, review), move the three notebook tickets out, and add tickets for the steps below.
A real name, description, provider, region, access level and last-updated date, plus the list of files behind each cohort. This fixes the generated names and unlocks the missing details on the catalog and cohort page, and the "only show what a cohort has" work.
The cart moves to the organization, then checkout and payment, then a "My purchases" page.
Together with Yahia's secure workspace: purchased data is delivered into the organization's secure workspace, and access is locked when the purchase expires.
Later: human review of requests, admin dashboard, provider portal and onboarding, notifications, analytics, plain-English search, comparison and export.