Decision Analytics For Parking Availability In Downtown Pittsburgh

Submitted By CiroFisicaro
Words: 4165
Pages: 17

Foundation of Operational Research/Business Analysis

Decision Analytics for Parking Availability in Downtown Pittsburgh

Individual assignment

Ciro Fisicaro

1. Introduction
This case study was located in Pennsylvania, in the Cultural District of Pittsburgh, core of business and economic activities of all south-western Pennsylvania. The Cultural District of Pittsburgh is exactly in the City centre, located in a 0.5 square miles called “Golden Triangle”, and as business centre, the Golden Triangle claim the presence of 130.000 workers that is the 32% of the total Pittsburgh’s workforce population. The traffic of people and vehicles in this zone has a relevant influence in Pittsburgh urban planning, due to inconvenient and congestions, and frequently the inconsistence of traffic service, as lacking of parking spots, damaged routes, damaged light service and imprecise signage, could enforce stress and pain for those people who use car as necessity. With the aim of implement an innovative solution for the parking problem, Pittsburgh Cultural Trust (PCT), a non-profit arts organization in Pittsburgh, established a collaboration with Traffic 21, a multidisciplinary
Carnagie Mallon University project, which objectives are to design, test, implement and finally select a new-tech solution in the transport sector.
The output between PCT and Traffic 21 collaboration generated a pilot programme live in December 2011 by a smart mobile application called ParkPGH, which provide a parking finder service and prediction of the spot availability with day in advance.

The nature and focus of the model
The nature of the model consist in a multivariable prediction model, based on quantitative and qualitative data collected from surveys, historical data, events data and other external resources (i.e. articles). The analysis concern on designing parking spot utilisation path and predict the available resources, car spots, verifying all the external variables (i.e. weather, events, period, etc.) that influence the variability of the spots availability and how this impact on the drivers’ perception about parking, taking care of Stakeholders desires and respecting limitations. The purpose of the model
The principal purpose of the predictive model was to develop an implemented system (ParkPGH app) able to predict with days in advance the percentage of future park availability in a particular parking Garage and show in real-time updates about parking spots, providing this service by different communication’ channels. The main purpose generated many different positive outcomes that affected the validity of the model: for example reduce the uncertainty of visitors and patrons on the event’s attendance through advance trip planning; on the other side, the model consistently help Garage management and owner, improving the optimization of parking utilisation and avoid bottle-neck when a garage is full.

Stakeholder analysis and achievements
Before develop an appropriate, affordable and clear model in order to collect and summarize all the necessary information, Traffic 21’ desires were to establish exactly who were the internal Stakeholders they were working for, listen to their requests and expectation and translate Stakeholders’ perceptions of the concern, respecting their point of view.

Team Traffic 21 built the model in line with stakeholder wants, which strongly depend on the perceptive context in which they operate, and created a needs assessment using surveys and historical data based on PCT/PCT patrons, Garage owners, Garage management and Pittsburgh Downtown Partnership, which characterised internal Stakeholders. Customers and visitors took part of the external Stakeholders groups as final application users, moreover due to the open source nature of ParkPGH, program’s Developers and Managers are external
Stakeholders as well. In fact this allow Developers to adopt the utilisation of the same model in other cities or implement the