I ran into a problem that: given a first name, I need to match the most likely email address from a given list of possible emails.

For example: given a first name like Mike, out of a list of all the email like binwang@datafireball.com, glenna@datafireball.com, mmarshall@datafireball.com, mikemarshall@datafireball.com…etc. I think mikemarshal@datafireball.com is the mostly likely one to be the correct match.

You can apply some algorithms to calculate the similarity between the first name and each email address. However, there are some situations where mmarshall@datafireball.com is better than binwang@datafireball.com and glenna@datafireball.com.. simply because of the tradition of use the first letter of first name in combination with last name. So on and so forth, simply using one algorithms might not work that well.

Since the amount of work is not too big to do manually… however, it is too big for ME manually. I decided to divide and conquer. So I broke each name mapping into a question and make it accessible through browser by using some basic php.

You can access the source code of the php from my Github repo, you might need to reverse engineer and guess how the MySQL table should look up based on your use case.

Here is a screen shot of how it looks like:

When the user submits the response, I not only document what his choice is, but also get the user agent, ip address and a cookie that I assigned to that user to track that user. So later down the road, some machine learning could be done to distinguish the irresponsible user from the informed and careful attendees.