New improvements to search and matching

+49 votes
762 views

Hi WikiTreers,

We just released a round of improvements, mostly to our search and matching functions.

---
1.) Secondary Sort.

As you may know, you are able to sort search results by first name, last name, birth date, death date, or profile manager.

Now you can also do a secondary sort, i.e. order the sorted results by first name, last name, birth date, death date, or profile manager.

For example, if you search for John Smith and you sort by first name, that will put all the Johns apart from those with John as a middle name. You can then order all the Johns by birth date or whatever you prefer.

Secondary sort has also been added to Watchlists (here's yours) and surname index pages (e.g. Smith).

Note that on surname index pages you have to break out of the default (cached) view to see all the sort options. Then the sort options are on the right.

This feature was requested here in G2G by Bob Keniston.

---
2.) Matching for "Unknown".

As some advanced WikiTreers have commented on here in the past, our search engine ignored people with the last name "Unknown". We were afraid that matching Unknowns with each other and with others would provide too many irrelevant results. Now that most profiles have dates this isn't as much of a danger.

I think the way we've done it now should be an improvement and help catch duplicates that were being missed before. That said, it may provide too many results in some cases. I'm especially concerned about MatchBot. I'm talking to the leaders of the MatchBot Monitors project about this. (That's a great and valuable project, by the way, and they could use more help here if anyone is interested.)

By the way, some members have requested that "Unknown" be searchable so that they can actually find people named Unknown. I appreciate the desire to do this, but it's still not possible. Unknown is treated as blank, i.e. "Unknown" is considered an empty last name, not as the last name Unknown. Searching for Mary Unknown, for example, will have as good a chance as matching Mary Smith or Mary Jones as matching someone named Mary Unknown.

---
3.) Improved Handling of Two-Word First Names.

Finally, we should now be doing a better job matching two-word first names. For example, "Marie A", "Marie A.", "Marie Anne", and "Marie-Anne" should now match with each other over other Maries and Annes.

We've also been working on improving the matching of two-word last names and last names with periods, hyphens, spaces, etc. Soon we should do a better job matching St. Laurent with St Laurent, St-Laurent, etc.

Onward and upward,

Chris

in The Tree House by Chris Whitten G2G Astronaut (1.5m points)
retagged by Ellen Smith

Hi Chris,

I am so thankful for all you do here. And so many others who spend hours to improve WikiTree. Bow!

Your work is just great!! Thank you so much. I am sure this will make it more easier for all of us.

I know you re spending lots of time but I still have a WikiTreeWish. And Santa will not fulfilI my wish. :/

I and I think many with me, would love a feature to search on city/town or at least country. That would help us so much at the Projects covering a specific country or region to find for example unsourced profiles or unconnected profiles of one country without scrolling long lists and canbe added fast to our Project so they will be looked after and improved. It also will help us I expect on finding out language for WikiTree Greeters and put the right Greeter or Mentor on the new-one..... 

To be honest I haven't the faintest how hard that is to make it possible. If so we will use the "Error Report"-way. 

So mabey I just have an impossible wish.

Thank you for reading.

Kind regards from the Netherlands.

Astrid Spaargaren,

Member of the Dutch_Roots_Project, WIkiTree Greeters and a BIG WikiTree lover.

Hi Astrid

Thank you for your comment; I agree with you that sorting by "Place of Birth" or "Place of Death" (those are the 2 geographical fields in the data base - what's in the biography is not possible for sorting) would be very helpful.  A problem might be that all "locations" are not written the exact same way on all profiles.  Perhaps a system using a "keyword" or "keywords" recognition would work - it might give one more "hits" than desired but still could limit some of those LONG lists significantly.

Chet Snow
Excellent! Thanx to all y"all.
Thank you Chet. Yes....... you re right. That is a problem. But the autofill nowadays at profiles might increase that problem. Or not........

Okay town or city not an option yet then. But country? The Netherlands might be spelled as Nederland (dutch way) Netherlands or Holland. Even variations with spelling mistakes up to ten variations is less work. Five entries instead of scrolling over thousands listed. :) That would be an improvement.

I really do understand the dilemma.... I will write Santa! :D

Thank you.

Astrid
Thank you Chris!  Really appreciate all the great improvements that have come about recently!  I just love WikiTree!!
I have been enjoying the search filters and I know that there must be a long list of desired improvements. Do you have a hit list to add in order? The one I find most useful in other sites is to be able to search for a person who has a known other (parent/child/spouse). From my examples, there are a huge number of Stephen Smith's - even when you limit it to a time-frame. There are very few married to Martha's and only one or 2 married to Martha Jones. If you are working on a One Name project, you rapidly build up repeated names and it really slows things down when you have to search through dozens of names to find the right one to add a child onto. Limiting it to your own watch list doesn't help much if you have entered 90% of the people with that name.

6 Answers

+19 votes
Thank you so much Chris and fellow WikiTreers!  The secondary sort will be so useful for names like "John Armstrong" or "Mary Smith" (both in my tree).  I'm sure there are many "Unknowns" that are matches to profiles already here.  Luckily, at least in Georgia, USA, DB errors are catching some of those with marriage to the same spouse or multiple marriages on the same day.

And, extra special kudos to Bob Keniston for expressing an improvement for all of us.
by Kathy Zipperer G2G6 Pilot (471k points)
+11 votes

May we please have a step by step example? e.g.

1. Go to https://www.wikitree.com/wiki/Special:SearchPerson

2. Search for First Name: John Last Name: Doe

3. Enter ??  to perform secondary sort.

I see a "Search again" option at the bottom of my search results but that may not be the new feature.

Thanks and sincerely, Peter

by Peter Roberts G2G6 Pilot (703k points)
edited by Peter Roberts

Peter:

Search:

1. Conduct a normal John Smith Search.

2. On the John Smith results page look for 

Sort by First Name | Last Name at Birth | Birth Date | Death Date | Profile Manager 
and then by First Name |Last Name at Birth |Birth Date |Death Date |Profile Manager 

At the top of the results. 

3. Click on the options to resort the results.

Watchlist.

1. Open your Watchlist.

2. Use the Sorting options shown at the top of your watchlist.
At the top of my Watchlist, the options are

Sort by Name | Birth Date  | Edit Date | Privacy  and then by Name | Edit Date | Privacy

I don't see an option to sort by death date (which would have been VERY useful last night as I was creating a list of people in my Watchlist who died in W W I).
Janet, make that suggestion as an improvement to WikiTree.
I  am confused.  I thought the Original Post said that sorting by death date has "also been added to Watchlists".

"Family List" is only a small subset of the Watchlist, and it doesn't have the option of sorting by death date either.

Hi Janet. I meant to say that secondary sort has been added to Watchlists, not that we added more sorting options to Watchlists. Watchlists and Surname Index Pages have certain constraints that we don't have on search results. We can be dealing with many thousands of items on these lists and only certain fields are indexed for quick sorting. That said, we could probably enable slow sorting by death date and profile manager, etc.

Thanks for the clarification.

'Slow" sorting on death date would probably still be faster than sorting on birth date and "sorting by eyeballs" on the death date.
+10 votes
Those are great improvements.  One possible suggest if it isn't being worked on already, is that when you start the two word first names, you include middle name as a searchable so that if you have say an existing  first name Mary and a middle name Ann, you can just click a box to combine them as first name and empty the middle name box.  Of course this ability will make people prone to combine names when they shouldn't, but it'd be useful when working on your own watch list.
by Dave Dardinger G2G6 Pilot (440k points)
+8 votes

I'm so relieved to see that people with LNAB of Unknown are no longer invisible to the WikiTree name search process! I know that I've started at least 3 G2G threads (possibly more) about that problem over the last 19 months. It's a great relief to confirm that I can finally find people like Anna van Londen (the profile focus of my first G2G on this subject) when I search for them by first name and current last name. Thanks to Chris and team!

I am also pleased to see the secondary sort feature, but after testing it I find that it's less wonderful than I was hoping it would be. It would be a lot more helpful if it were also possible to (1) see a second screen of search results, instead of just the first 100 results, and (2) search in reverse order (for example, sort by first name, starting with the letter "Z").

Two more improvements I'm still hoping to see sooner instead of later.

  • Ability to use the wildcard character "?" in a search. For example, if I'm looking for a woman whose name might be spelled Elizabeth or Elisabeth, it would be very nice to be able to search for Eli?abeth. Currently I can find her by searching for Eli*, which also returns people name Eli, Elisha, Elijah, etc. (often returning well over 100 results) or I can do two searches. A wild card character would save a lot of trouble.
  • I'm very much looking forward to the ability to see two-word last names be treated as the equivalent of the same name that doesn't contain a space. As I said last September in a G2G thread I started regarding needed improvements in search, please teach the search function to overlook the underscore character _ in last names like Van_Dyke and Du_Bois and van_der_Walt when searching for possible duplicates. Currently, if I'm looking to see whether there's an existing profile for an Abraham Van Arnhem, I can enter Abra* and Vanarnhem in the search boxes, and I'll see results for Abram or Abraham with last names of Vanarnhem, Vanaernum, Vanarnum, VanAernam, VanArnem, and more. However, if I want to find variant spellings for the two-word version of the name (and this was properly a two-word name), I have to search separately for Abra* with each possible variant spelling (Van_Arnhem, Van_Aernum, van_Aernam, van_Arnheim, etc.). How hard would it be to teach the search function to treat Van_Arnhem as equivalent to VanArnhem, for the purpose of finding possible duplicates?
by Ellen Smith G2G Astronaut (1.5m points)
Hi Ellen. I started a new thread to discuss whether to ignore spaces in last names: https://www.wikitree.com/g2g/402174/should-spaces-last-names-ignored-for-most-purposes-wikitree
+5 votes
Chris, thanks for these improvements. Could you please explain how the Unknown search has been improved? An example would be great. Thanks.
by Jillaine Smith G2G6 Pilot (907k points)

I'm not Chris, but I can give you an example, Jillaine.

Do a name search for Agnes Curtis. Before this change, Agnes Curtis of Nazeing, Essex, died 1585, wouldn't show up in the results list because her LNAB was unknown. Now she's there. HURRAH!!

So basically,we still can't search for Unknown's, but we can search on a current last name (assuming it's there), and those who have Unknown in LNAB will be included on the list?

Just wanting to make sure I'm clear...

Name search also can find Unknowns now by a name in the "Other Last Names" field. I confirmed this by searching for "Mary Symonds" and verifying that Mary (Unknown) Larabee was on the results list.

But we still can't find "Mary Unknown" using the WikiTree name search utility.

However, a serious WikiTree user who knows a "Mary Unknown" profile exists can find that profile by a Google search for:

site:Wikitree.com mary formerly surname unknown [plus other keywords such as the name of the town]

+4 votes
Chris I can tell you many ways how Wiki tree is be hind the eight and must Improve, I am currently debating if I want to stay or leave, I can put on your boards, the Pros and Cons, weakness and things that need to improve or you will very left in the dust forever.  I hope you can make wiki a better place for every one.
by S Sagers G2G6 Mach 2 (28.8k points)

Related questions

+37 votes
25 answers
+23 votes
2 answers
+4 votes
2 answers
141 views asked Feb 11, 2021 in WikiTree Tech by Martin Lillieborg G2G4 (4.7k points)
+20 votes
2 answers
267 views asked Nov 29, 2017 in WikiTree Tech by Ros Haywood G2G Astronaut (1.9m points)
+11 votes
2 answers
277 views asked Jun 27, 2016 in WikiTree Tech by S Willson G2G6 Pilot (223k points)
+147 votes
27 answers
4.6k views asked Sep 4, 2018 in The Tree House by Chris Whitten G2G Astronaut (1.5m points)
+9 votes
0 answers
+6 votes
1 answer
+26 votes
2 answers

WikiTree  ~  About  ~  Help Help  ~  Search Person Search  ~  Surname:

disclaimer - terms - copyright

...