TraBina

How to use TraBina

TraBina turns an event log into a process map. You bring cases, activities, and times — from a file, database, HTTP API, or a short Python script. The map shows how work actually flows, where it waits, and which paths are common.

1. Create a process and upload a log

Open New process, name it, then pick a source on the Upload screen:

  • File — CSV or Excel.
  • Database — MySQL, PostgreSQL, SQL Server, or Oracle. Direct host/port and SSH tunnels are both supported.
  • API — call a JSON HTTP API (GET, POST, and the usual auth options).
  • Python — fetch and reshape nested JSON into flat rows for mining.

Every source needs a case id (one ticket or order), an activity (the step name), and at least one time column. A second end time is better when you have it (Disco-style Start + End). A resource column (person or team) is optional.

If the preview has only one time column, choose it for Start time and leave End time on “no, my file has only one time column”. The map still works; it cannot tell how long each step was worked, only the gap until the next step. If column names are not guessed, map them on the mapper screen, then mine. Persian headers are supported.

Database, API, and Python can also save a schedule (once, repeating interval, or days of the month). Import history on Upload lists every versioned log — who imported it, when, and a download.

2. API

Use the API tab for a standard JSON response that is already a flat array of objects (or an array under a short path). Flow: Test the call → Preview rows → Map case / activity / time → Import as log.

Set method and URL, then Auth (none, bearer, basic, or API key), Params, Headers, and Body as needed. Tokens stay encrypted. If the array is nested, set a JSON path under Records (for example data.items). Leave the path blank when the response is already an array, or a common wrapper like data / results / items.

Private and loopback hosts are blocked. For deeply nested or multi-level payloads, switch to the Python tab instead.

3. Python

Prefer Python when the payload is nested or you need to reshape fields (like a Power BI custom connector). Assign a list of dicts to result, or return that list from main(). Use http.get / http.post for outbound calls — the same private-host rules as API apply.

Same flow as API: Test → Preview → Map → Import. The editor starts with a small sample; commented lines show how to walk nested lists into case, activity, and time fields. Only the process owner can edit scripts; editors can pick a saved script and import.

4. Read the map

Each box is an activity. Each arrow means “this step is followed by that one.” Line thickness is how often that path happens. Color follows the metric you pick (frequency, case frequency, or duration).

Duration on a box is how long that step was worked (end − start). Duration on an arrow is idle time after the step finished until the next one started. If start and end are the same, boxes show 0 and all time sits on the arrows. Secondary adds a second number on the box; it does not recolor the map.

Use the Activities and Paths sliders to hide rare steps. Happy path highlights the most common variant. Find jumps to an activity by name.

5. Filters

Tick variants in the right list, or set dates and extra attributes. The violet box counts how many cases would remain. That count updates as you change the filter; it does not rebuild the map yet.

Apply filters remine the log with those rules. Clear filters goes back to the full log. Map, analysis, cases, resources, stats, log, and export all use the applied filter.

6. Saved views and copy

Open Saved views & copy at the bottom of the filter pane. Save this view stores the current filter on this process. Open applies a saved view. Copy to a new process makes a separate process with the same log and that filter already applied. The copy appears in My space; this page stays where you are.

7. Other tabs

  • Analysis — slow activities, slow handoffs (waiting between steps), and rework (a step that repeats in the same case).
  • Cases — each case’s cycle time (first start to last end). Click a case for its timeline.
  • Resources — who hands work to whom, if you mapped a resource column.
  • Stats — case duration histogram, events per day, and a variant Pareto. Bars on the Pareto are case counts; the copper line is the cumulative share of cases covered by the most common paths.
  • Log — a preview of the filtered events.
  • Conformance — check the log against a BPMN model if an editor uploaded one.

8. Export

On the map, the Export group has PNG and SVG (the map image) and PDF (an A4 report: map, KPIs, happy path, variants, activities, and rework). Export CSV in the page header downloads the filtered event log.

9. Team and account

After mining finishes and the process is Ready, the owner can invite people who already have a TraBina account. They keep their own workspace; they only gain access to this process. You can replace the log (file, database, API, or Python), or delete the process or your account from Account.