Clean Up Your Family Tree: Remove Duplicates, Fix Dates and Standardise Places

Cleaning a family tree in the wrong order means doing the same work twice. A practical order:

1. Encoding first. If accented names are mangled, every search and every comparison from this point forward will work on the wrong strings. Fix the encoding before you fix anything else.
2. Duplicates next. Merging duplicates after fixing the records inside them means your fixes get duplicated. Find and merge first; clean afterwards.
3. Structural errors. Broken parent-child links, families pointing to nonexistent IDs — these distort every downstream report and have to be fixed before sourcing or sharpening makes sense.
4. Dates and places. Run the Vital Sharpener and Location Standardiser to tighten and normalise vital facts.
5. Sourcing last. Once the data is clean, attaching citations is far less likely to need redoing.

Re-export your file and re-run GEDminer after each step. The Tree Health Score will tell you, in a single number, whether your work moved the needle.

Place names are messy because they evolve: counties merge, parishes are absorbed, countries change name and borders move. The Location Standardiser does not try to "correct" history — it groups variants of the same place so you can pick a canonical form for each cluster.

Variants are grouped by:

• Component matching — comma-separated location pieces are matched piece by piece, so "Manchester, England" matches "Manchester, Lancashire, England".
• Case and accent folding — "Müller" matches "Mueller" matches "MULLER".
• Common abbreviations — "Yorks" → "Yorkshire", "Co. Cork" → "County Cork".
• Ordering tolerance — "England, Manchester" matches "Manchester, England".

Each cluster shows the count of records using each variant, so you can pick the canonical form (usually the longest, fully-qualified version) and apply it back in your editor.

Not every vague date is worth fixing. A vague death date for a peripheral cousin matters less than a vague birth date for a great-great-grandparent who anchors a whole branch.

The Vital Sharpener scores every imprecise vital fact by:

• Centrality — how many descendants are downstream of this person in your tree?
• Recency — more recent vague dates are usually easier to research and produce more downstream improvement.
• Type — birth dates rank above marriage dates which rank above death dates.
• Source potential — facts in eras with strong record coverage (e.g. 1841+ in England) are worth fixing first because the records exist.

The result is a prioritised work list. Spend an hour at the top of the list and your tree's quality score will jump noticeably; spend it at the bottom and the change is barely measurable.

Adding citations is slow because every fact needs a real source attached. But sourcing is what turns a tree from a private collection into a research-grade asset that other people can verify.

GEDminer measures sourcing as a single percentage: total citations divided by total facts, expressed as a percentage. A tree at 0% has no sources. A tree at 100% has at least one citation per fact. Most published, research-grade trees land in the 50–80% range.

The cleanup pattern that works:

• start by attaching citations to vital facts (birth, marriage, death) of every direct ancestor,
• then to vital facts of every collateral ancestor within four generations,
• then to non-vital facts (occupation, residence, immigration) for direct ancestors,
• and only then to peripheral relatives.

This pattern produces the largest jump in the sourcing score for the smallest investment of time.

Re-export and re-analyse after each significant cleanup pass — after a duplicate merge session, after a place-name standardisation pass, after a sourcing sprint. Comparing scores between exports is the only objective way to know whether your work is producing measurable improvement.

A useful rhythm: end every research session with a quick re-export and a glance at the score components. Set a target (e.g. "I want sourcing above 60% by the end of the year") and let the score guide your effort allocation between the cleanup categories.