Extract
Reads the pages you have already scraped and keeps only the genuine civil-engineering opportunities. Nothing is downloaded again, so you can change your keywords and run this as often as you like.
Try a different threshold…
Two measures, each with its own threshold. Score counts matched terms using the points you give each dictionary — by default only the civil-engineering one, so score 0 means nothing civil matched. Merit is the weighted total: every match adds the weight beside it in its dictionary, scaled by that dictionary’s own rule, so it rises with how much matched as well as how strongly, and goes negative when only noise matched. Drag the slider to see how the blocks below would be judged. Nothing is saved until you press Save, and nothing is re-downloaded. Change what each dictionary contributes →
What happened
- Pages processed
- 0
- Blocks read
- 0
- Passed the filters
- 0
- New opportunities
- 0
- Already known
- 0
Why content was dropped
New opportunities
What was scraped, and what happened to it
| Verdict | Content found on the page | Score / Merit | From | Why it was dropped |
|---|---|---|---|---|
| Nothing selected yet. | ||||
Advanced & tuningkeyword audit, custom conditions, re-runs, tidy-up, clearing a source
Keywords doing the sorting
Every discipline keyword, how many items it actually brought in, and a one-click way to drop one that is pulling in noise.
My conditions
Write your own tests over the blocks in the table above — score, keywords, dictionary type or drop reason — and see what they would catch.
Tidy up after a rule change
Re-running adds and refreshes, but never removes: items kept under looser rules stay in the feed. This deletes the ones that no longer pass. Starred items are left alone, and nothing is tombstoned — loosen the rule again and they come straight back.
Re-run on everything
Applies your current rules to every page already downloaded. Nothing is fetched again, and anything that now passes is added. Use this after changing your keywords.
Clear a source’s data
Deletes every page, block and opportunity produced by the source picked above, so the next scrape of it starts clean.