proposal.Rmd

---
title: 'Proposal: an online tool for cycle policy planning and public engagement'
author: "Dr. Robin Lovelace (University of Leeds), Dr. Nikolai Berkoff (independent web developer), Martin Lucas-Smith (CycleStreets Ltd)"
date: "16 February, 2015"
output:
  pdf_document:
    keep_tex: yes
  word_document: default
---

```{r, echo=FALSE, results='hide'}
pkgs <- c("png", "grid")
lapply(pkgs, library, character.only = TRUE)
```


# Introduction

The proposal is to create a tool which will empower
decision-makers, stakeholders and others to understand cycling at the local level
and to focus on the areas and routes most in need of improved cycle infrastructure.

The concept of this planning tool was developed during the HackMyRoute
hackathon organised by the Leeds node of the Open Data Initiative (ODI).
Representatives from the ODI, West Yorkshire Combined Authority (WYCA)
and members of the public were enthused by the possibilities it could open up.
The proposed work will extend the functionality and usability of the tool.

Our aim simple: to create an online
and mobile application that will provide an evidence-base for understanding
variability in the rate of cycling at the local level across West Yorkshire.
The tool will provide up-to-date information on a range of factors affecting cycling
and sustainable transport policy to transport planers, policy makers, and
interested members of the public in an accessible, attractive and interactive form.

```{r, echo=FALSE}
# Todo: add link to the wicid website here
# Removed:
#   The tool could also help decision-makers implement more cost-effective pro-cycling
# interventions
# harnessing the power of online, interactive map visualisation
# technology.
```

# Objectives

- To create an interactive map that covers the entirety of West Yorkshire,
with information on the current rate of bicycle commuting in different areas.
- To enable users to visualise the rate of cycle commuting, not only in specific areas,
but also along 'flow lines'.
- To create metrics of the potential level of cycling in different areas, based
on the number of cyclable trips between areas on the map.
- To allocate flow lines of particular interest to the user onto the transport
network using the CycleStreets.net API.(e.g. where the current
highest rate of cycling is, or where the potential increase in cycling is highest)
- To allow users to visualise additional information relevant to cycle policy
in the local area including bicycle parking, road traffic casualties and
bicycle shops. This will help priorities what type of cycling provision is
most needed in different areas.

# Example usage

The tool could be used by policy makers to identify new paths that serve areas with currently
low cycling rates but which have a high potential for increased cycling.
Many people currently have access to the underlying data, but not presented in
a form that is easy to understand.  Thus policy makers could use the tool to
tap into existing open data to create an evidence base to support
better decision-making.

```{r, echo=FALSE}
# another proposed path may be in an area with an already
# medium cycle rate but have larger propensity to cycle rate with the
# connecting location.
## By decision makers
## By cycle campaigners

```

Another beneficiary will be cycle campaigns. Local campaign groups
can easily become focussed on single issues, to the detriment of
understanding the 'big picture' and rates of cycling and provision in
areas outside their members' local areas. The tool could help campaigners
prioritise where to focus their next campaign, empowered with the evidence.

The public and everyday cyclist will benefit by being provided with
a simple, cohesive system that makes available a wide range of evidence on
cycling in a single place. This will inform the ongoing cycling debate in
West Yorkshire, directing it to discussion of the data rather than anecdote.

```{r, echo=FALSE}
# This would not simply be using currently available data as there may be
# many other metrics to which users might like to have access.  The tool
# would also have the ability to create their own custom metric based on
# the available data. For example, one metric could combine data on the
# current fastest and safest route along with the difference the new
# infrastructure would make.  We propose to create this map for Leeds.
```

# Proposed work

We will design, build and deploy this online map-based tool.
The map display will have a simple interface allowing the user to
select datasets of interest using drop-down menus on both cycling
behavior and the transport infrastructure. Choropleth maps,
flow lines and markers will
provide an overview of many different factors quickly and
simply. When the user has found routes that are of interest,
he or she will be able to plot the current safest route for cyclists
and to see where it could be made more direct.

We will use 'agile development' to prioritise the workflow. Based
on real-time user feedback and testing, the tool will undergo
continual and rapid refinement. The open-source nature of the
tool will ensure that the end result will not be a static piece
of software but a collaborative process between the developers,
planners and other users.

The project will build-on and complement existing research
activity being undertaken at the University of Leeds, and with
data layer-based work by CycleStreets Ltd (who run the UK cycle
journey planner at cyclestreets.net). If it commissions this
work, West Yorkshire could be an early adopter of a powerful tool
that could eventually be deployed nationwide.

# Why is this different to current software?

The 'open data' revolution has greatly increased access to data. This
can benefit transport planners, if they have access to appropriate and
user-friendly tools, built on a strong evidence base and accessible to
public scrutiny.

Propensity to cycle, for example, is a critical metric in deciding where
to target new interventions and can be calculated by harnessing the flow
data and census databases mentioned previously.

Extracting useful *information* and *knowledge* from diverse datasets
requires specialist skills to analyse. This is a major barrier to entry
for transport planners, policy makers and stakeholders from civil
society. In addition, new datasets continue to emerge that could aid
transport planner; this tool will be flexible enough to rapidly incorporate
these new sources of data.

# Prototyping work

A proof-of-concept was developed during the
HackMyRoute hackathon referred to above. 'FixMyPath' illustrates the current
level of cycling across Leeds, based on Census commuting flow data.
Collision data, route accessibility,
and cycle parking are also available to users. The prototype can be seen at tested at
https://robinlovelace.shinyapps.io/fixMyPath/. An example of what the tool will
look-like, when deployed in a smartphone app, is illustrated below. The fact
we have created a working prototype in one day of intensive work demonstrates
that we have the ability to deliver a finished product in a short period of time.

```{r, echo=FALSE}
# I think people know what cyclestreets.net is! (RL)
# The partners for the proposal include CycleStreets Ltd, a not-for-profit
# social enterprise who run a cycle journey planner and a range of tools
# for advocacy groups and local authorities. Amongst these is a
# comprehensive collision data viewer, for instance. Funding will enable
# collaboration between the partners in the project so that the combined
# strengths of the (a) research-based activity which will underpin the
# tool, and (b) the end-user reach and user knowledge available to
# CycleStreets will mean a user-friendly and widely-accessible tool.
```


```{r, echo=FALSE, fig.height=4}
img <- readPNG("../../mobile-layer-example.png")
grid.raster(img)
```


# Finances

The proposed work will cost in the region of £5,000 to £40,000 depending on the
features and specification of the tool:

- The £5,000 option will extend the existing tool across all of West Yorkshire and
allow the user to select the origin-destination flows of interest based on multiple criteria.
- £12,000 will pay for the previous step and deployment of the tool in a
simple prototype smartphone app to encourage user engagement, testing and feedback.
- £20,000 will pay for the previous steps and additional work to allow for
the display of propensity to cycle between all MSOA zones
(containing on average 7,000 people) in West Yorkshire on desktop computers.
The additional money
will be spent on modelling work to estimate propensity to cycle under a range of
scenarios and the visualisation of tens of thousands of flow lines on an interactive map.
- £40,000 will pay for the full flow-based model to be deployed in a user-friendly
and functional smartphone app. A key feature of this app will be to allow user
input, based on interpretation of the aforementioned data, on the routes that
should be prioritised for future upgrades to cycle paths. This will allow for
'crowd-sourcing' of data to help transport planners decide where to focus limited
resources across West Yorkshire. Money will be set-aside for outreach and
engagement events to ensure high uptake of the app.

# Contacting us

We are willing to discuss the details of this plan and are willing to
negotiate a plan that is of greatest benefit for the funders.
Please contact us at info@cyclestreets.net to let us know the
outcome of this proposal or for further information/discussion.


```{r, echo=FALSE}
# Appendix: data sources

# The data used for this project is publicly available. We selected
# datasets based on their ability to help explain the high or low rate of
# cycling in different areas, and to guide future active travel policies,
# such as the construction of new bicycle paths and dedicated bicycle
# parking facilities. Specifically, we used the following datasets:
#
# - Data on travel to work flows between medium super output areas
# (MSOAs) --- 'origin-destination flow data'. This data has recently (summer 2014)
# been made available to the public through the Open Government Licence (OGL):  .
# There are 107 MSOA zones that comprise the Local Authority of Leeds.
# The input dataset could easily be expanded to incorporate all 300 MSOA zones
# of the West Yorkshire Combined Authority (WYCA), including
# Leeds, Calderdale, Bradford, Kirklees and Wakefield.
# For Leeds, on which our prototype was based, the flow data contains 10,536
# origin-destination pairs. An example of this data is present in Table 1.
#
# ```{r, echo=FALSE}
# # library(knitr)
# # kable(fl_present[1:5, ])
#
# ```
#
# |Origin    |Destination | Total| Bicycle| lon_origin| lat_origin|  lon_dest|
# |:---------|:-----------|-----:|-------:|----------:|----------:|---------:|
# |E02002330 |E02002330   |    66|       1|  -1.400577|   53.92915| -1.400577|
# |E02002330 |E02002331   |   742|      38|  -1.400577|   53.92915| -1.347653|
# |E02002330 |E02002332   |     2|       0|  -1.400577|   53.92915| -1.705267|
# |E02002330 |E02002333   |     9|       2|  -1.400577|   53.92915| -1.687695|
# |E02002330 |E02002334   |    50|       1|  -1.400577|   53.92915| -1.357652|

```