Deliverables
      1) The scoring tool: How likely will they adopt a kid?
      Please enter the profile of an applicant and get the score!
      2) The matching tool: How to present the family with the kind of kid profile they prefer?
      Please enter the profile of a foster candidate and get Top 3 types of children for matching up!
Limitation
    1) More information other than demographic statistics about foster families are in need to build a decision tree with higher accuracy of prediction (around 73% currently).
    2) The frequency distribution of different ethnic families is quite skewed, which may erode the generability of the prediction model. 
References: 
Github Link: https://github.com/ChuhengHu/FosterCare