Any logical reason why you "disagree whole heartedly?" Also if you download the app and click the info icon it explains how the data is being collected. I just don't want to share that on a reddit post is all.

If I share the exact dataset that will literally be a recipe for someone to build a similar app, especially if I do that in a reddit post. If you want, you can email me at streetwisesafe@gmail.com and we can chat over email there and I can be more specific. Also every claim that has presented I have refuted, so the app is definitely both well-intentioned and feasible.

You are the only person who has asked to see the actual data source and cannot trust it without knowing that, not "most here."

From an api that performs web scraping on several datasets across the US. Based on the testing I have done the crimes are extremely recent, from within the last day to the past week, and new ones are inputted each day that can change the recommended route. I don't feel comfortable sharing the exact source as that might give too much info away for competitors.

Also if you don't trust the reporting/data source you can still use the app to see the locations of nearby crimes, and the user is not restricted to take one route, he has a plethora of options to optimize his preferences for both safety and speed.

Not using BPD

1. The user has the option to choose between a plethora of options based on his preferences for both speed and safety, they are not just confined to take one specific route.

2. With or without the app people will always be killed. I also state in my disclaimer that the app only gives you the lowest chance of a crime happening and does not guarantee that one will not happen. The safety is still the responsibility of the user.

3. I have done extensive testing with the database and it actually returns very recent crimes (within the last day) up till the last week. It is also constantly updated hourly with new crimes that can change the recommended route based on the algorithm.

Hence, I disagree with all of your points.

This is a good point, but as I stated previously, current crime hotspots do indeed have predictive value of where future crimes will occur. Moreover, I am not forcing the user to take a specific route, they have a plethora of options to choose based on their preferences for both speed and safety. If you don't believe me you don't have to download the app.

Also your whole comparison of Baltimore to NYC is completely dumbfounded because the user will be choosing between routes that go through the same city, with all route options having being taken a similar number of times. Hence if one route has a significantly greater number of crime hotspots that will almost certainly correlate with the crime:trips attempted ratio. Also as I said previously, the user can choose between multiple options and I am not forcing him to take any specific route.

Edit: brightbehaviorist I totally understand what you're saying that the crime:trips ratio would be a better indicator of the safety than just crimes, and I am working on integrating this into the algo. However, I think it is still very useful to see local crime hotspots along the route you're taking, and I has said before, the user can choose between a plethora of options based on their preferences (like google maps). Thanks for your feedback!

I did extensive market research before pursuing this idea, and seeing all the studies on the correlation is what inspired me to build it. In fact similar apps also use a similar principle to calculate areas where future crimes are likely to occur. An example of such an app is WalkSafe+. It's your choice if you want to believe that.

You're right the app cannot predict whether someone will have an incident, but it definitely lowers the chance a considerable amount. Hence "some predictive value" is most certainly different from "no predictive at all." It's your choice if you want to download the app though and don't think it's benefitting you. I don't think you speak for everyone though.

1. People using GPS definitely use the visual side as a guide, but to say they are glued to their phones is definitely an exaggeration.

2. This is true, there are definitely some people who know the streets very well and don't need the know the safest route. However, to say most people who walk home at night fall in this category is also an exaggeration.

3. But what if the person doesn't want to waste money on an Uber?

4. I think this app could have a lot of use for college students walking to parties late at night or little kids who are walking home from school.

5. I think the app will definitely help those walking at night feel safer knowing the locations of nearby crimes.

6. I also thought about this while building the app. The user has the ability to choose between a plethora of options that maximize their preferences for safety and speed, they are not just being forced to choose one route. Moreover, the routes are all calculated from where the user is located, so different users will get different routes depending on their individual locations in the city. Lastly, the database is getting updated hourly with new data from the past day that could change the recommended routes, so if a criminal catches someone it would most likely be due to random chance rather than because the app is making the user more vulnerable.

This is a fair point, black users could be more subject to police brutality, but like you said, that doesn't fall into the category of crimes. The safety the app calculates is solely related to crimes, and I state in my disclaimer that every user is still responsible for his or her safety/life at the end of the day. The user also has the option to choose between a plethora of route options that optimize his preferences for both safety and speed, they are not being forced to take just one route.

Many studies have shown that past crimes are indicative of future crimes happening at similar locations in the future, which is what inspired me to build the app. Hence the model definitely has some predictive value and is not obsolete at all, imo.

I am literally using a database of real-time verified crime data that gets updated hourly with new, very recent crime reports that gets factored into the algorithm to determine the safest route. How is that not "backed up by data"? You are also right that a lot of the time the safety score for each route are very similar, which is why the user has the option to choose between different routes (they are not just being forced to take one route) and pick one that optimizes their preferences for both safety and speed.

I never understood the whole "racist argument" actually. If your safety is in peril, why would you consider socio-ethical ramifications of your choices? I've seen some mentally unstable people brandishing objects that could be used as weapons on the street. Would I take into consideration that they are possibly unfortunate victims of an unjust society (which I agree with) and walk by within the distance of their reach? For someone of a small stature like myself, I'd probably put the concern of safety over anything like that, to be honest.

I thought about this while building the app. After the app calculates the safest route the gps navigation takes over (just like google maps), and the user is free to look at his surroundings instead of constantly having to look at his phone. Maybe you should download the app and try it yourself if you have any more questions.

As I stated in my last reply, the safest route is being calculated by comparing each route to each other, not by examining specific neighborhoods. As long as the route does not exclusively fall in a neighborhood where no crimes are being reported, a reasonable safety score can be calculated. I agree that there are some crimes that are not being reported, but no app can reliably get such data since it is unreported. I'm also working on adding a feature where users can manually report local crimes and then factor that into the safety algorithm. It's your choice if you want to download the app though, but there are definitely many people who could benefit from it.

I would only see this being an issue if the user is traveling between multiple neighborhoods that cross into those underreported areas. The safest route is calculated by compared the relative safety of each route to each other based on proximity to local crimes, so if the app is used in neighborhoods with a decent reporting rate I don't think it would be disingenuous at all. The safest route is not being calculated by comparing different neighborhoods to each other, rather the routes that are all within one neighborhood.

You're right, there is definitely a lot of crime that goes unreported, however unreported crimes cannot be accessed from any sort of database so making such an app using unreported crimes would basically be impossible. Also while a lot of crime is unreported, that does not mean that using reported crimes would be entirely inaccurate as there is a lot of recent crime that does get reported in the database as I have done extensive testing with the app. Thanks for your feedback!

It uses machine learning to calculate local crime hotspots for recent crime based on police reports from a database and then assigns a danger score to each route based on the proximity of the route to nearby hotspots. The safest route with the lowest danger score is the default route the app highlights.

It's not four years old lol, it's meant for people who are ages 4 or above is what that means. That one review probably meant it has helped him up to this point. It has only been on the app store for about a month. Sorry for any confusion!