A Jira Project, Untangled

Every story, task and bug is a dot. Its date sets where it sits from left to right, under the bars of the sprints, and its team sets its row. The bigger the dot, the more story points it carries. Parent links tie each issue to its feature, the features to their epics and the epics to the one initiative at the top. Blocking links run between the rows, so when one team waits on another, you see the line.

The comments hang off the issues they belong to, replies off the comments they answer. A decision made in a thread sits next to the work it changed.

The chains of blocked work end at one task: the staff pilot go/no-go on 4 November. Six items block it, and only one of them, the feature flag, is done. Still open are the regression suite, the privacy sign-off, the penetration test, the load test and, since 14 October, a reopened bug.

Every issue carries how many open blockers it has and whether another team holds them, so one filter shows all the work that is waiting: 13 issues, and for every one of them the blocker belongs to another team.

Three teams' work waits on one backend story, the card transaction event stream: the risk score, the guardian's push message and the spending categories. It was carried over from Sprint 3 when the card processor's test feed was down for two days, and on 13 October the product owner raised it to Highest. Behind it sits everything in spending limits and insights.

The map shows the same shape once before. The per-child sub-ledger held up three items, the top-up, the reconciliation job and the allowance schedule, until it was done on 24 September; then the chain moved on. One bottleneck cleared, one still open, side by side.

The child app needs its own app ID, signing certificates and push keys. Nobody planned that work. It surfaced on 1 October as a question in a comment, who owns the push key?, and was filed the next day. It still sits in the backlog with no owner and no estimate, yet it blocks two stories planned before it: the child app's PIN sign-in in Sprint 4 and the push message in Sprint 5.

It is not the only blocker planned too late. Five issues are marked as waiting on work scheduled after them, the go/no-go among them: its load test is planned for Sprint 6, after the decision.

QA filed it on 9 October: when the clocks go back an hour, a child's allowance is paid at 06:00 and again an hour later. It was fixed and closed on 13 October and reopened the next morning, because the catch-up job that pays missed runs keeps its own copy of the schedule.

Its connections show how far it reaches. It was caused by the allowance schedule API, it now blocks the set-allowance screen and the pilot go/no-go, the nightly reconciliation job should catch it, and a second report on 15 October was closed as its duplicate. Its five comments tell the story in order, and it must be fixed by 23 October, before the clocks change on the 25th.

A Jira project already has the parts: issues, their parents, their blockers, sprints, teams and comments. Make each issue type a node type, keep the links as connections, and turn the fields you want to compare by, such as status, team, story points or flagged, into properties you can filter on. See how properties and types work, or ask us for help preparing the data.