Local Mobility Index (LMI)
A Methodological Exploration
As a newcomer to Lesvos island, I am keen to learn about the different places and how to get to them. However, the only source of information I could find is this website: https://www.ktel-lesvou.gr/index.php?option=dromologia, which I find too complex to grasp.
The lecture by Prof. Thanasis and Prof. Spilanis on the value of developing an index was inspiring to me because it can measure states, and changes, to generate meaningful comparisons. What fascinates me is not just the process of calculating established indices, but also how to conceptualize, measure, and create a new index.
“Let’s build a new index then!” I thought.
Through this writing, I intend to quantify the difficulties or easiness of a destination on Lesvos Island based on its accessibility and affordability which I call Lesvos (or Local for more universal use) Mobility Index/LMI.
Local Mobility Index (LMI)
What is Local Mobility Index (LMI)?
LMI is about assessing how easy it is to get around on Lesvos Island (by Bus because this is the only data I have now). To better understand the LMI, there are a few key concepts to know:
- Mobility refers to the ability to get from one place to another
- Accessibility is about how easy it is to reach a certain location
- Connectivity is the state of being connected
- Proximity nearness in space, time, or relationship
- Affordability ability to be afforded/inexpensiveness
How do we define “how easy” to get around?
When it comes to defining how easy it is to get around on Lesvos island, there are a few criteria that I think would be helpful to consider:
- Connectivity: The number of bus lines serving the location.
- Trip frequency: The availability of bus services and how frequently they run (measured in total trips per week both from and to Mytilene with each location).
- Spatial proximity: The distance from Mytilene (in kilometers).
- Temporal proximity: The travel time from Mytilene (in minutes).
- Affordability: The price of a bus ticket (transportation services) from/to Mytilene (measured in Euro).
In the Impact Assessment and Policy Evaluation class, Prof. Frans Sijtsma stressed the importance of avoiding over-articulation and under-articulation of measurement (or preferences). To adhere to this principle, I aim to maintain the monetary valuations as monetary and the non-monetary valuations as a composite index. Therefore, the conceptual calculation model is as follows:
The Local Mobility Index (LMI) can be defined as the total amount of euros required to achieve one accessibility score of a particular location. In other words, it can also be referred to as the Cost-Accessibility Index.
How to Construct this Index?
Collecting the Data
I started the groundwork by collecting data from the KTEL website. I recorded every line, every stop, and their attributes such as distance, total trips per week, duration, and ticket prices, and supplemented it with spatial information, including the approximate latitude and longitude of the bus stop. The resulting table looks like this:
Interpolating the Missing Values
However, some data was missing, so I used a tool called Inverse Distance Weighting in ArcGIS Pro to estimate those values. I transformed the data into a point shapefile and used IDW interpolation to create a raster for each attribute. Then, I filled in the missing values using the Extract Multivalues to Point tool and exported the final data to a CSV file for further analysis. Interestingly, by using this IDW, we can also produce maps based on ticket costs that must be paid per trip such as “fare zones”. The following are the maps of the interpolated results:
Normalizing the Values
Since non-monetary variables have different units of measurement (number of routes and total weekly trip in number, distance in km, travel time in minutes), we cannot directly aggregate them before applying a formula to normalize their values. For frequency and connectivity, higher values are more desirable. Conversely, for spatial and temporal proximity, lower values are better. Thus, we need to adjust their values to the range of 0 to 1 using different formulas as follows:
Creating a Composite Index
Once we have adjusted the values of non-monetary variables using the formula for normalization, we get a standard scale for each element, even though the value is no longer an absolute figure but a relative one. To create a composite index, I choose to simply add up all the indices. This results in the lowest Accessibility Score being 0 and the highest Accessibility Score being 4, just like a 4-point GPA system.
Here are the Accessibility Scores for all the towns/bus stops) on Lesvos:
On average, small bus stops have higher accessibility scores compared to large bus stops and have less dispersion/variation of values (shorter interquartile range). Large bus stops show almost a normal distribution without any outliers, while small bus stops show positive skewness with one outlier identified: Lampou Milli.
What about the monetary variable? We keep it in its original form, as a monetary value in Euros. Generally speaking, larger bus stops on Lesvos Island tend to have higher travel costs compared to smaller bus stops. This trend is consistent with larger stops exhibiting a positively skewed distribution of travel costs and smaller stops showing a negatively skewed distribution. However, we did not observe any significant outliers in either stop type. Additionally, smaller stops appear to be slightly more dispersed in terms of their travel costs than larger stops (longer interquartile range).
Measuring Local Mobility Index (Cost-Accessibility)
As a result, we now have two numbers for each stop. The first number represents the ease of reaching the stop, and it’s calculated by adding up the route availability and diversity, weekly frequency, distance, and travel time. A higher number means it’s easier to reach the stop, and a lower number means it’s harder to reach the stop. The second number represents the monetary cost of traveling to the stop (from or to Mytilene) in Euros. By dividing the monetary value and non-monetary value we can get the “cost-accessibility index”
Here are the Accessibility Scores for all the towns/bus stops) on Lesvos:
Reflection: Why LMI is a not-so-good index?
Although the Local Mobility Index (LMI) can condense a lot of information about a destination’s accessibility and affordability into a single number, it is not without its shortcomings. For instance, some of the variables used in the index might be redundant, such as route availability and trip frequency. It’s likely that if a stop has more routes, then there will be more trips as well. Additionally, factors like distance and travel time are also quite similar except when it is impacted by other factors, such as infrastructure quality, topography, or road layout (I guess this is the case of Argennos?). Moreover, the cost of travel might also be a function of distance and travel time, which could lead to severe multicollinearity. In other words, the variables might be so interrelated that it’s challenging to distinguish the individual effects of each factor.
Furthermore, the LMI is incomplete because it only captures the bus as a mode of travel due to data limitations. It ignores other aspects of mobility, such as driving (by car, scooter/motorcycle), biking, or walking, which can also impact a destination’s accessibility and affordability (but then it will become too complex too). Another limitation is that it may be somewhat groundless, as it lacks a strong theoretical foundation (Obviously! Since this LMI thing just popped out of my curiosity).
Lastly, even if we know that Sigri has the highest LMI (the most expensive destination in terms of a cost-accessibility index), it’s not clear what we should do about it. Since the distance variable is fixed, we can’t move Sigri closer to Mytilene. Therefore, the only way to improve the LMI is by increasing the weekly frequency, lowering the travel cost, and lowering the travel time by introducing other modes or technology, or adding more routes. However, we don’t know if the islanders (both in Sigri and Mytilene) or tourists need these changes, making the relevance (and benefit) of the LMI somewhat unclear.
Who Cares?
While the Local Mobility Index (LMI) can be a useful tool for researchers and policymakers to understand the accessibility and affordability of different destinations, sometimes, people only want to know what’s most important to them: which bus lines they need to take if they want to go somewhere. (For a more high-res version of the map, visit this link).
In the end, the usability of the index depends on its ability to provide valuable insights that can be translated into actionable steps to drive change and improve the current situation.
Try again!
Luthfi Muhamad Iqbal
Mytilene, 23/02/2023
