Social Needs Marketplace - Impact

Project Lead: Dr. Daniela Rosu

Project Team:
 
Contributors: 

2020 
Graduate Students
Theo Shiyuan Lin (CS)
Undergraduate Students 
Lester Dishu Lyu (CS)
Sheilla Shojaie (CS)
Lawrence Ling Feng Zhao (CS) 

2019 
Undergraduate Students 
Simon André (MIE)
Stephen Gidge (MIE)
Brandon Lista (MIE)​
​Sheilla Shojaie (CS)
Theo Shiyuan Lin (CS)
Theresa Taylor (MIE)
Fred Yuzhu Yao (CS) 
Dawas Zaidi (CS) 

2018 
Undergraduate Students 
Shivam Batt (CS) 
Maitreyee Sidhaye (EngSci)
Betty Runtian Wang (CS)
Lisa Tzu-Chiang Weng (EngSci)

2017 
Leon Lukashevsky (Web Developer)
Undergrad Students 
Lucas Arcuri (MIE)
Angad Chadav (MIE)
Aneel Sood (MIE)
Tom Nguyen (MIE) 

2016 
Graduate Students 
Chiang Liu
Yi Ru
Undergraduate Students
Audrey Fang (CS)
Otto Fang Lin (CS) 
Flora Qiu Liu (CS) 


Recent Publications






 
 
The Problem
 
Despite Canada being a wealthy country, there are many communities that have unmet needs. Immigrants, refugees, indigenous, and impoverished individuals often lack the financial support, social support, and/or mobility to access needed social services. This problem persists despite government efforts, NGOs, and other organizations that provide goods and services to those in need. A major cause of the problem is the complexity of the system for matching and delivering goods and services to the people who would benefit most. Both the unique needs of individuals as well as the vast number of organizations that may potentially satisfy their needs is at the heart of the problem.

Matching social services to client needs is currently a time-consuming and tedious process. Social workers must manually search for services and goods in their organization’s internal files or over the Internet. Unorganized and scattered information, as well as the extensive evaluation of information significantly reduces the efficiency and effectiveness of meeting client needs. To address this issue, the Centre for Social Services Engineering has developed.

Project Description

The Social Needs Marketplace-Impact (SNM-I) assists agencies in efficiently and effectively servicing the needs of their clients. SNM-I provides a market infrastructure that enables the matching of demand to supply. Addressing the privacy concerns of agencies, SNM-I provides a private market in which a single agency is able to manage their clients’ needs (i.e., goods and services), and match them with the goods and services the agency provides, as well as those provided by individual donors affiliated with the agency.  The matching process uses AI constraint satisfaction techniques to ensure that potential matches satisfy all relevant, spatial, temporal, product, recipient, and other constraints on both the demand and the supply side. Addressing another aspect of market efficiency, an agency’s private market can be expanded, if an agency has a partnering relationship with other agencies that are willing to share portions of their supply side, namely goods and services that they deem shareable, thus creating an overlap in their private markets. Furthermore, agencies can include in their private market, goods and services provided by third party service providers, such as furniture banks and ESL training.

SNM-I has been deployed and evaluated with the help of social agencies in Toronto, who offered feedback and suggestions for improvement. The information collected during this evaluation has been used to improve and expand the platform and will be disseminated via MIE technical reports and papers published in relevant journals and conferences. The revised application is being deployed and will be continuously evaluated by the participating organizations.  


Goals
 
  • Develop an efficient way to redistribute new or used goods and services by allowing the “demand” side to clearly identify their needs and the “supply” side to make known what is available​
  • Match goods and services provided by donors to the clients’ individual needs
  • Address the logistical issue of transporting goods and/or services from the supplier to the consumer within an acceptable range of time. The Marketplace supports a volunteer/agency supported by a logistical system that uses an Uber-like volunteer network
  • Enable more efficient use of NGOs by virtually combining and focusing their resources to achieve better outcomes​
  • Allow partial, controlled access to collaborating agencies’ and communities' private markets so that they can share goods and services
 
 
Objectives
 
There are five objectives that are addressed by the Social Needs Marketplace:
 
1. Learn user profiles.  It is essential to learn the goods and services an individual requires. Initial individual user profiles are built by converting self-declarations (i.e., statements in English made by supply-side users that reveal what they are willing to supply) into a semantic representation. Using machine learning, SNM-I can learn the categories of clients, and many of their needs. These categories can be used to infer needs that the client fails to self-declare.
 
2. Represent knowledge of users, goods, services, logistics, preferences, etc.  There are three main challenges involved with describing goods and services. First, the terminology that people use for these descriptions is rife with ambiguity. Second, it is often unclear whether or not the terminology used is adequate to completely describe the range of goods and services. Third,  many of the terms that people use to describe goods and services are ad hoc and arbitrary. SNM-I has developed an ontology of client needs and  goods and services that enables the precise specification and matching of goods and services to client needs.

3. Allow Goods & Services to be shared across agencies. SNM-I incorporates both a local and integrated marketplace. Agencies and communities that have goods and services (and donors), that they may wish to share with other agencies/communities can designate which goods and services and the agencies they wish to share them with.  For the designated agencies the selected goods and services are visible to their version of SNM-I and are included in the matching to their clients’ needs.

4. Increase the effectiveness of meeting client needs. It is important to match client specific needs (demand) to available goods and services (supply). To evaluate effectiveness, client needs must not only be matched but they also must be satisfied. Since matching of client needs will be done across a constantly growing market of goods and services, greater efficacy will only be achieved as more needs are met.
 
5. Overcome the challenge of physically transporting the donated goods from the pickup location to the delivery location. Creating an online marketplace provides a more efficient means of redistributing new and used goods and services but there remains logistical challenges that are addressed by SNM-I:
 
  • How do we use the online marketplace to match donated transportation services with requested pickup-and-delivery tasks?
  • How do we address the challenges of (1) geography and time (i.e., the driver will need to physically meet with both the supplier and consumer at convenient times without going far out of his/her way), (2) incentives and trust (i.e., how we establish a system where suppliers and consumers will be comfortable and safe in interacting with volunteer drivers), and (3) demand planning, prediction, and market clearance (i.e., how do we set up a system that will ensure that the required pickup-and-deliveries can be performed)?
 
 
Using SNM-I
 
In the supply side of the system (Figure 1), each provider would provide one or more goods or services with constraints on their availability. All goods and services to be donated can be added interactively or via a CSV file. Real Time availability for both goods and services is displayed on the map for Case Managers to see.​​

 
Figure 1: Suppliers Point of View
 
 
​​
 

In the demand side of the system (Figure 2), social workers and case managers can access all the clients entered into the system. They can easily add, delete, or modify clients and their information. Each client has a client profile which contains personal information. This profile includes specifications of their individual needs (i.e. language of service, distance to service location, time of day, etc.). Client needs are categorized as either a Goods Needs or Service’s Needs. Clients can have any number of goods and services they require. These needs are identified and recorded by social workers or client managers during client assessments.

​Figure 2: Clients Point of View


SNM-I automatically matches client’s needs to available goods and services (Figure 3). However, it is also possible to manually search for goods and services as well.

Figure 3: Matching Needs to Goods & Services

Finally, the status of a need can be updated whether it has been contacted, matched, or referred (Figure 4).  The status will then be reflected in the interface for others to be aware of. 

Figure 4: Updating Status of Need

Liu, C., Aleman, D.M., & Beck, J.C., " Modeling and Solving the Senior Transportation Problem ," Proceedings of the Fifteenth International Conference on the Integration of Constraint Programming, Artificial Intelligence, and Operations Research (CPAIOR2018), 412-428, 2018.
Rosu, D., Aleman, D.M., Beck, J.C., Chignell, M., Consens, M., Fox, M.S., Gruninger, M., Liu, C., Ru, Y., & Sanner, S., " A Virtual Marketplace for Goods and Services for People with Social Needs ," Proceedings of the IEEE Canada International Humanitarian Technology Conference 2017, 201-206, 2017.                                                                                    
Liu, C., Aleman, D.M., & Beck, J.C., " The Senior Transportation Problem ," AAAI Workshop on AI and OR for Social Good, San Francisco, U.S., February 2017.
Howe, L., and Fox, M.S., (2017), " Increasing the Effectiveness of the Non-Profit Sector Through Visualization: A Case Stufy of Furniture Banks ," Proceedings of the 2017 Industrial and Systems Engineering Conference, K. Coperich, E. Cudley, H. Nembhard, eds.
Fox, M.S., (2018), “ The Semantics of Populations: A City Indicator Perspective ”, Journal of Web Semantics, Vol. 48, pp. 48-65.
Rosu, D., Aleman, D.M., Beck, J.C., Chignell, M., Consens, M.P., Fox, M.S., Gruninger, M., Liu, C., Ru, Y., Sanner, S., (2017), “Knowledge-Based Provisioning of Goods and Services: Towards a Virtual Marketplace”, Proceedings of the AAAI 2017 Symposium on AI for Social Good, Technical Report SS-17-01.