Post by kickinrocks on Apr 4, 2018 17:58:31 GMT
Hi all,
I'm a software developer by trade and spent some time a few summers ago working on some genealogy research with some relatives of mine. We have had this mysterious missing link named Grandpa Ron in our family for many years. He disappeared from my dad's life before he was born and held the last remaining link to that side of the family.
His name was very common, so common that an ancestry.ca search done using his name, approximate date of birth and localized to all of Canada yielded us nearly 10 thousand records.
So my cousins and I set out to find Grandpa Ron and I made us a tool to help collaborate on our research. Here is a screenshot of a more modernized version of what we called RONDB:
We would cherry pick our leads from the list of ancestry.ca records and trace their whereabouts all over Canada. Adding snippets of information like census locations, highschool photos, adoption records etc. This then allowed me to rate our list of "suspects" against our "master profile". The "master profile" contained all of the information we knew about Grandpa Ron. Like the town he lived in when he met my grandma and that his mothers first name was Grace and a few other small details. Some of the facts that we had were likely false, we knew Grandpa Ron was a con artist and likely didn't give my Grandma an entirely true story about his background. This is why we thought the ranking system worked best.
After 8 months of data input by 4 people, we had 700 profiles and almost 3000 bits of additional information. Our ranking system highlighted 10 or so "suspects" that very closely fit our master profile. We decided at that point to start cold calling our top 10, or at least family members of those 10 to see if we could find a match. Dozens of phone calls later and we came up with nothing.
I was a bit dejected at that point and shelved the project for quite a while, nearly a year actually. In the meantime one of my cousins was adding more suspects and stumbled upon an interesting character on ancestry.ca that had somehow been linked to 5 or 6 family trees through broken marriages. He didn't really fit our profile and had used a different last name a few times. He was in our database twice as two separate profiles that happened to share two crossing points. They both briefly lived in New Liskeard, ON and then in the same town my Grandma met Ron in.
I called in a huge favor to a local law enforcement friend, who ran a last known address search on the most recent name and address we had for this guy. It was from 1984, but it was worth a shot. And wouldn't you know it, we got him. He had been dead for almost 10 years at the time, but he had a family and two wives in the area that allowed us to get the closure we needed.
Anyway, the point of all this is, I still have the code from RONDB and have been dusting it off the last few weeks since my interest in this case was peaked again. Would anyone else think its worth it to start something like this for EARONS? A place where we could collaborate on POIs, add evidence etc. and maybe work out some kind of scoring system to help us narrow the search?
Obviously there would be some privacy concerns by storing peoples names, but I think I could work around those by not displaying the names on the profile to anyone that hasn't a) logged in and b) contributed to that profile in some way. Another feature off the top of my head would be to merge duplicate profiles, to allow for better collaboration(with the option to unmerge if you aren't satisfied with the match). This would search for similar profiles based on a set of criteria and then merge them into one, similar to what Wikipedia does with duplicate entries.
Let me hear what you think.
I'm a software developer by trade and spent some time a few summers ago working on some genealogy research with some relatives of mine. We have had this mysterious missing link named Grandpa Ron in our family for many years. He disappeared from my dad's life before he was born and held the last remaining link to that side of the family.
His name was very common, so common that an ancestry.ca search done using his name, approximate date of birth and localized to all of Canada yielded us nearly 10 thousand records.
So my cousins and I set out to find Grandpa Ron and I made us a tool to help collaborate on our research. Here is a screenshot of a more modernized version of what we called RONDB:
We would cherry pick our leads from the list of ancestry.ca records and trace their whereabouts all over Canada. Adding snippets of information like census locations, highschool photos, adoption records etc. This then allowed me to rate our list of "suspects" against our "master profile". The "master profile" contained all of the information we knew about Grandpa Ron. Like the town he lived in when he met my grandma and that his mothers first name was Grace and a few other small details. Some of the facts that we had were likely false, we knew Grandpa Ron was a con artist and likely didn't give my Grandma an entirely true story about his background. This is why we thought the ranking system worked best.
After 8 months of data input by 4 people, we had 700 profiles and almost 3000 bits of additional information. Our ranking system highlighted 10 or so "suspects" that very closely fit our master profile. We decided at that point to start cold calling our top 10, or at least family members of those 10 to see if we could find a match. Dozens of phone calls later and we came up with nothing.
I was a bit dejected at that point and shelved the project for quite a while, nearly a year actually. In the meantime one of my cousins was adding more suspects and stumbled upon an interesting character on ancestry.ca that had somehow been linked to 5 or 6 family trees through broken marriages. He didn't really fit our profile and had used a different last name a few times. He was in our database twice as two separate profiles that happened to share two crossing points. They both briefly lived in New Liskeard, ON and then in the same town my Grandma met Ron in.
I called in a huge favor to a local law enforcement friend, who ran a last known address search on the most recent name and address we had for this guy. It was from 1984, but it was worth a shot. And wouldn't you know it, we got him. He had been dead for almost 10 years at the time, but he had a family and two wives in the area that allowed us to get the closure we needed.
Anyway, the point of all this is, I still have the code from RONDB and have been dusting it off the last few weeks since my interest in this case was peaked again. Would anyone else think its worth it to start something like this for EARONS? A place where we could collaborate on POIs, add evidence etc. and maybe work out some kind of scoring system to help us narrow the search?
Obviously there would be some privacy concerns by storing peoples names, but I think I could work around those by not displaying the names on the profile to anyone that hasn't a) logged in and b) contributed to that profile in some way. Another feature off the top of my head would be to merge duplicate profiles, to allow for better collaboration(with the option to unmerge if you aren't satisfied with the match). This would search for similar profiles based on a set of criteria and then merge them into one, similar to what Wikipedia does with duplicate entries.
Let me hear what you think.