More frequently we hear that algorithms are used in organizations to optimize profits. This can be seen in many banks, where the best econometricians and data scientists are hired to make optimal processes just a fraction more profitable.
In today’s society, however, we see that there is an increase in appreciation for people working in the service sector. It is therefore questionable whether it is wise to use the best data scientists in our society to make the rich even richer. This while there are still many unsolved problems which impact the entire society.
Deploying algorithms in the social sector
Algorithms have a great potential, and they are already able to offer help in the service sector. For example, they can support the police in detecting criminals, relieve the burden on healthcare workers by automatically detecting tumors and help teachers to support their students.
Use algorithms at child welfare organizations
Child helplines are one of the children’s welfare organizations where algorithms can provide help. Today we see that it is difficult for child helplines to find enough suitable volunteers who have adequate knowledge and enough time to answer messages from children in need. Furthermore, we see that many questions from children return regularly and that the work of child helplines is therefore not as challenging as it could have been.
Classify incoming messages
Incoming messages from the child helplines can be classified, for example based on neglect. This can be done by using natural language processing (NLP). This is a technique that can extract relevant information from texts and thereby understand human language. This allows children with standard questions to be helped with the answers to previous messages from the database. This leaves time to help children who need immediate help.
Problems of text classification
Natural language processing is a technique that has been around for years and has proven to be valuable in classifying certain texts. Still, it is important to be aware that the texts that child helplines receive are often of a different nature. These are texts of children (in distress) that may contain typos, incorrect sentence structures, abbreviations and chat language. It is also important that the algorithm works ethically. For example, it is important that girls are treated just like boys and that factors such as writing style, religion and origin should not play a role in the classification of the letter. In short, there are plenty of problems that smart econometricians and data scientists can address. But then we first need to know how to value, challenge and excite these people to work in the social sector.
Sharona Boonman (23 years old) is a graduate researcher at the University of California at Los Angeles (UCLA). She works with anonymized data from a child helpline with the aim of providing neglected children with the help they deserve. This research is part of her Erasmus Mundus joint master’s degree, a prestigious, integrated, international study program, in digital communication technology in Europe.
