In this article we address the study of parking occupancy data published by the Birmingham city council with the aim of testing several prediction strategies (polynomial fitting, Fourier series, k-means clustering, and time series) and analyzing their results. We have used cross validation to train the predictors and then tested them on unseen occupancy data. Additionally, we present a web page prototype to visualize the current and historical parking data on a map, allowing users to consult the occupancy rate forecast to satisfy their parking needs up to one day in advance. We think that the combination of accurate intelligent techniques plus final user services for citizens is the direction to follow for knowledge-based real smart cities.
http://dx.doi.org/10.1007/978-3-319-59513-9_11
Pulmonary drug delivery system M.pharm -2nd sem P'ceutics
Predicting Car Park Occupancy Rates in Smart Cities
1. PREDICTING CAR PARK OCCUPANCY RATES
IN SMART CITIES
Daniel H. Stolfi1
dhstolfi@lcc.uma.es
Enrique Alba1
eat@lcc.uma.es
Xin Yao2
x.yao@cs.bham.ac.uk
1Departamento de Lenguajes y Ciencias de la Computación,
University of Malaga, Spain
2CERCIA, School of Computer Science,
University of Birmingham, Birmingham, U.K.
International Conference on Smart Cities
Smart-CT 2017
Málaga, Spain
June 14-16 2017
2. CONTENTS
1 INTRODUCTION
2 OUR PROPOSAL
3 EXPERIMENTS
4 CONCLUSIONS & FUTURE WORK
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 1 / 21
3. CONTENTS
1 INTRODUCTION
2 OUR PROPOSAL
3 EXPERIMENTS
4 CONCLUSIONS & FUTURE WORK
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 1 / 21
4. CONTENTS
1 INTRODUCTION
2 OUR PROPOSAL
3 EXPERIMENTS
4 CONCLUSIONS & FUTURE WORK
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 1 / 21
5. CONTENTS
1 INTRODUCTION
2 OUR PROPOSAL
3 EXPERIMENTS
4 CONCLUSIONS & FUTURE WORK
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 1 / 21
7. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Road Traffic
Parking in a Big City
INTRODUCTION
Nowadays in our cities. . .
There is a larger number of vehicles in the streets
Finding an available parking space is hard
Time and Fuel are wasted in finding a free space
Tons of greenhouse gases are emitted to the
atmosphere
The citizens’ quality of life is decreasing
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 2 / 21
8. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Road Traffic
Parking in a Big City
INTRODUCTION
Nowadays in our cities. . .
There is a larger number of vehicles in the streets
Finding an available parking space is hard
Time and Fuel are wasted in finding a free space
Tons of greenhouse gases are emitted to the
atmosphere
The citizens’ quality of life is decreasing
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 2 / 21
9. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Road Traffic
Parking in a Big City
INTRODUCTION
Nowadays in our cities. . .
There is a larger number of vehicles in the streets
Finding an available parking space is hard
Time and Fuel are wasted in finding a free space
Tons of greenhouse gases are emitted to the
atmosphere
The citizens’ quality of life is decreasing
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 2 / 21
10. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Road Traffic
Parking in a Big City
INTRODUCTION
Nowadays in our cities. . .
There is a larger number of vehicles in the streets
Finding an available parking space is hard
Time and Fuel are wasted in finding a free space
Tons of greenhouse gases are emitted to the
atmosphere
The citizens’ quality of life is decreasing
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 2 / 21
11. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Road Traffic
Parking in a Big City
INTRODUCTION
Nowadays in our cities. . .
There is a larger number of vehicles in the streets
Finding an available parking space is hard
Time and Fuel are wasted in finding a free space
Tons of greenhouse gases are emitted to the
atmosphere
The citizens’ quality of life is decreasing
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 2 / 21
12. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Road Traffic
Parking in a Big City
INTRODUCTION
Nowadays in our cities. . .
There is a larger number of vehicles in the streets
Finding an available parking space is hard
Time and Fuel are wasted in finding a free space
Tons of greenhouse gases are emitted to the
atmosphere
The citizens’ quality of life is decreasing
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 2 / 21
14. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Road Traffic
Parking in a Big City
SENSORS
Sensors reporting car park occupancy
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 3 / 21
15. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Predicting Car Park Occupancy Rates
Case Study: Birmingham U.K.
Predictors Analyzed
SYSTEM ARCHITECTURE
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 4 / 21
16. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Predicting Car Park Occupancy Rates
Case Study: Birmingham U.K.
Predictors Analyzed
BIRMINGHAM, U.K.
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 5 / 21
17. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Predicting Car Park Occupancy Rates
Case Study: Birmingham U.K.
Predictors Analyzed
BIRMINGHAM, U.K.
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 5 / 21
18. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Predicting Car Park Occupancy Rates
Case Study: Birmingham U.K.
Predictors Analyzed
BIRMINGHAM, U.K.
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 5 / 21
19. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Predicting Car Park Occupancy Rates
Case Study: Birmingham U.K.
Predictors Analyzed
BIRMINGHAM, U.K.
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 5 / 21
20. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Predicting Car Park Occupancy Rates
Case Study: Birmingham U.K.
Predictors Analyzed
BIRMINGHAM, U.K.
32 Car Parks32 Car Parks
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 5 / 21
21. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Predicting Car Park Occupancy Rates
Case Study: Birmingham U.K.
Predictors Analyzed
DATA SOURCE
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 6 / 21
22. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Predicting Car Park Occupancy Rates
Case Study: Birmingham U.K.
Predictors Analyzed
DATA SOURCE
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 6 / 21
23. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Predicting Car Park Occupancy Rates
Case Study: Birmingham U.K.
Predictors Analyzed
DATA SOURCE
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 6 / 21
24. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Predicting Car Park Occupancy Rates
Case Study: Birmingham U.K.
Predictors Analyzed
DATA SOURCE
Data set:
Oct 4th to Dec 19th
From 9am to 5pm
18 measures per day
32 car parks
36,285 occupancy measures
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 6 / 21
25. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Predicting Car Park Occupancy Rates
Case Study: Birmingham U.K.
Predictors Analyzed
PREDICTORS
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 7 / 21
26. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Predicting Car Park Occupancy Rates
Case Study: Birmingham U.K.
Predictors Analyzed
POLYNOMIALS, FOURIER, TIME SERIES & K-MEANS
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 8 / 21
27. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Predicting Car Park Occupancy Rates
Case Study: Birmingham U.K.
Predictors Analyzed
POLYNOMIALS, FOURIER, TIME SERIES & K-MEANS
Polynomial Fitting
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 8 / 21
28. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Predicting Car Park Occupancy Rates
Case Study: Birmingham U.K.
Predictors Analyzed
POLYNOMIALS, FOURIER, TIME SERIES & K-MEANS
Polynomial Fitting
Fourier Series
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 8 / 21
29. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Predicting Car Park Occupancy Rates
Case Study: Birmingham U.K.
Predictors Analyzed
POLYNOMIALS, FOURIER, TIME SERIES & K-MEANS
Polynomial Fitting
Fourier Series
Time Series
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 8 / 21
30. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Predicting Car Park Occupancy Rates
Case Study: Birmingham U.K.
Predictors Analyzed
POLYNOMIALS, FOURIER, TIME SERIES & K-MEANS
Polynomial Fitting
Fourier Series
Time Series
K-Means
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 8 / 21
31. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Predicting Car Park Occupancy Rates
Case Study: Birmingham U.K.
Predictors Analyzed
KM-POLYNOMIALS AND SHIFT & PHASE
F(x) = a0 + a1 · x + a2 · x2
+ . . . + an · xn
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 9 / 21
32. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Predicting Car Park Occupancy Rates
Case Study: Birmingham U.K.
Predictors Analyzed
KM-POLYNOMIALS AND SHIFT & PHASE
F(x) = a0 + a1 · x + a2 · x2
+ . . . + an · xn
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 9 / 21
33. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Predicting Car Park Occupancy Rates
Case Study: Birmingham U.K.
Predictors Analyzed
KM-POLYNOMIALS AND SHIFT & PHASE
F(x) = a0 + a1 · x + a2 · x2
+ . . . + an · xn
F(x) =(a0 + δ) + a1 · (x + φ) + a2 · (x + φ)2
+ . . . + an · (x + φ)n
δ = Shift, φ = Phase
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 9 / 21
34. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Predicting Car Park Occupancy Rates
Case Study: Birmingham U.K.
Predictors Analyzed
KM-POLYNOMIALS AND SHIFT & PHASE
F(x) = a0 + a1 · x + a2 · x2
+ . . . + an · xn
F(x) =(a0 + δ) + a1 · (x + φ) + a2 · (x + φ)2
+ . . . + an · (x + φ)n
δ = Shift, φ = Phase
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 9 / 21
35. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Predicting Car Park Occupancy Rates
Case Study: Birmingham U.K.
Predictors Analyzed
KM-POLYNOMIALS AND SHIFT & PHASE
F(x) = a0 + a1 · x + a2 · x2
+ . . . + an · xn
F(x) =(a0 + δ) + a1 · (x + φ) + a2 · (x + φ)2
+ . . . + an · (x + φ)n
δ = Shift, φ = Phase
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 9 / 21
36. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Predicting Car Park Occupancy Rates
Case Study: Birmingham U.K.
Predictors Analyzed
KM-POLYNOMIALS AND SHIFT & PHASE
F(x) = a0 + a1 · x + a2 · x2
+ . . . + an · xn
F(x) =(a0 + δ) + a1 · (x + φ) + a2 · (x + φ)2
+ . . . + an · (x + φ)n
δ = Shift, φ = Phase
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 9 / 21
37. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Predicting Car Park Occupancy Rates
Case Study: Birmingham U.K.
Predictors Analyzed
KM-POLYNOMIALS AND SHIFT & PHASE
F(x) = a0 + a1 · x + a2 · x2
+ . . . + an · xn
F(x) =(a0 + δ) + a1 · (x + φ) + a2 · (x + φ)2
+ . . . + an · (x + φ)n
δ = Shift, φ = Phase
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 9 / 21
39. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Training
Predicting
Prototype
K-FOLD CROSS VALIDATION
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 11 / 21
40. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Training
Predicting
Prototype
K-FOLD CROSS VALIDATION
K=10
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 11 / 21
41. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Training
Predicting
Prototype
MEAN SQUARED ERROR (MSE)
MSE = 1
n i(yi − fi)2
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 12 / 21
42. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Training
Predicting
Prototype
K-FOLD TRAINING RESULTS (I)
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 13 / 21
43. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Training
Predicting
Prototype
K-FOLD TRAINING RESULTS (I)
Polynomial Fitting
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 13 / 21
44. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Training
Predicting
Prototype
K-FOLD TRAINING RESULTS (I)
Polynomial Fitting
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 13 / 21
45. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Training
Predicting
Prototype
K-FOLD TRAINING RESULTS (I)
Polynomial Fitting Fourier Series
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 13 / 21
46. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Training
Predicting
Prototype
K-FOLD TRAINING RESULTS (I)
Polynomial Fitting Fourier Series
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 13 / 21
47. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Training
Predicting
Prototype
K-FOLD TRAINING RESULTS (II)
K-Means
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 14 / 21
48. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Training
Predicting
Prototype
K-FOLD TRAINING RESULTS (II)
K-Means
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 14 / 21
49. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Training
Predicting
Prototype
K-FOLD TRAINING RESULTS (II)
K-Means KM-Polynomials
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 14 / 21
50. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Training
Predicting
Prototype
K-FOLD TRAINING RESULTS (II)
K-Means KM-Polynomials
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 14 / 21
51. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Training
Predicting
Prototype
SHIFT & PHASE AND TIME SERIES
Shift & Phase
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 15 / 21
52. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Training
Predicting
Prototype
SHIFT & PHASE AND TIME SERIES
Shift & Phase
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 15 / 21
53. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Training
Predicting
Prototype
SHIFT & PHASE AND TIME SERIES
Shift & Phase Time Series
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 15 / 21
54. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Training
Predicting
Prototype
SHIFT & PHASE AND TIME SERIES
Shift & Phase Time Series
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 15 / 21
55. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Training
Predicting
Prototype
PREDICTION RESULTS: UNSEEN WEEK
Which car park will be best for me
tomorrow?
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 16 / 21
56. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Training
Predicting
Prototype
PREDICTION RESULTS: UNSEEN WEEK
Which car park will be best for me
tomorrow?
And the day after tomorrow?
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 16 / 21
57. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Training
Predicting
Prototype
PREDICTION RESULTS: UNSEEN WEEK
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 16 / 21
58. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Training
Predicting
Prototype
PREDICTION RESULTS: UNSEEN WEEK
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 16 / 21
59. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Training
Predicting
Prototype
PREDICTION RESULTS: UNSEEN WEEK
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 16 / 21
61. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Training
Predicting
Prototype
PREDICTION EXAMPLES
Working Days
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 17 / 21
62. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Training
Predicting
Prototype
PREDICTION EXAMPLES
Working Days Weekends
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 17 / 21
63. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Training
Predicting
Prototype
PARKING IN BIRMINGHAM
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 18 / 21
66. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Conclusions
Future Work
CONCLUSIONS
Six predictor for forecasting car park occupancy rates
Trained by using real parking data
Tested with one week of unseen parking data
Time Series shows the best results during working days
Shift & Phase has good results during weekends
We have presented a web based prototype
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 20 / 21
67. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Conclusions
Future Work
CONCLUSIONS
Six predictor for forecasting car park occupancy rates
Trained by using real parking data
Tested with one week of unseen parking data
Time Series shows the best results during working days
Shift & Phase has good results during weekends
We have presented a web based prototype
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 20 / 21
68. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Conclusions
Future Work
CONCLUSIONS
Six predictor for forecasting car park occupancy rates
Trained by using real parking data
Tested with one week of unseen parking data
Time Series shows the best results during working days
Shift & Phase has good results during weekends
We have presented a web based prototype
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 20 / 21
69. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Conclusions
Future Work
CONCLUSIONS
Six predictor for forecasting car park occupancy rates
Trained by using real parking data
Tested with one week of unseen parking data
Time Series shows the best results during working days
Shift & Phase has good results during weekends
We have presented a web based prototype
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 20 / 21
70. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Conclusions
Future Work
CONCLUSIONS
Six predictor for forecasting car park occupancy rates
Trained by using real parking data
Tested with one week of unseen parking data
Time Series shows the best results during working days
Shift & Phase has good results during weekends
We have presented a web based prototype
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 20 / 21
71. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Conclusions
Future Work
CONCLUSIONS
Six predictor for forecasting car park occupancy rates
Trained by using real parking data
Tested with one week of unseen parking data
Time Series shows the best results during working days
Shift & Phase has good results during weekends
We have presented a web based prototype
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 20 / 21
73. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Conclusions
Future Work
FUTURE WORK
Repeat this study using a larger training data set and
other cities
Include new predictors in the comparison
Develop an application for mobile phones
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 21 / 21
74. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Conclusions
Future Work
FUTURE WORK
Repeat this study using a larger training data set and
other cities
Include new predictors in the comparison
Develop an application for mobile phones
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 21 / 21
75. Introduction
Our Proposal
Experiments
Conclusions & Future Work
Conclusions
Future Work
FUTURE WORK
Repeat this study using a larger training data set and
other cities
Include new predictors in the comparison
Develop an application for mobile phones
Daniel H. Stolfi, Enrique Alba & Xin Yao Predicting Car Park Occupancy Rates in SC 21 / 21
76. QUESTIONS
Predicting Car Park Occupancy Rates in Smart Cities
Prototype: http://mallba3.lcc.uma.es/parking/
Questions?
Daniel H. Stolfi
dhstolfi@lcc.uma.es
Enrique Alba
eat@lcc.uma.es
Xin Yao http://neo.lcc.uma.es
x.yao@cs.bham.ac.uk http://danielstolfi.com
Acknowledgements: This research has been partially funded by Spanish MINECO project TIN2014-57341-R (moveON). Daniel H. Stolfi is
supported by a FPU grant (FPU13/00954) from the Spanish Ministry of Education, Culture and Sports. University of Malaga. International
Campus of Excellence Andalucia TECH.
77.
78. PARAMETERIZATION
Training days: Oct 4th to Dec 12th
Testing Week: Dec 13th to Dec 19th
Predictor Parameter Training
Polynomials: 2o Degree Fold: 1
Fourier Series: 3 Components Fold: 1
K-Means: 3 Clusters Fold: 1
KM-Polynomials: 2o Degree Fold: 1
Shift & Phase : - Fold: 1
Time Series : - Weeks: 8