In this week’s In-Ear Insights, Katie and Chris discuss corporate social responsibility, social good, and ways of making a difference that don’t necessarily involve large financial contributions to causes. Listen in for an upcoming example of how Trust Insights will be using its powers for good.
Sponsor: Knowledge Matters
In-Ear Insights is brought to you by Knowledge Matters.
Do you teach marketing at the collegiate level? Imagine using immersive, visual, interactive marketing simulations designed for today’s digitally-native students.
Besides bringing marketing concepts to life, it’s cloud based and includes automatic grading and easy integration with all the popular collegiate learning management systems such as Canvas, Blackboard or Moodle.
If you would like to check out case simulations that cover all the key concepts for an Intro to Marketing course like Pricing, Promotion, Market Research, Market Segmentation, and much more, sign up today for a demo at: www.knowledgematters.com/podcast
Listen to the audio here:
- Need help with your company’s data and analytics? Let us know!
- Join our free Slack group for marketers interested in analytics!
Machine-Generated Transcript
What follows is an AI-generated transcript. The transcript may contain errors and is not a substitute for listening to the episode.
This is In-Ear Insights the trust insights podcast.
In this week’s episode of In-Ear Insights, we are talking about social good corporate social responsibility as it relates to data science and machine learning. So Katie, let’s start out by talking about what are what are some of the big overarching responsibilities that a company in our space has towards this community? What How should people be thinking about corporate social responsibility? You know,
it’s such an important topic. And I think that people get overwhelmed by the idea of a social good branch of their company, because you automatically think, well, I don’t necessarily have money to donate or, you know, I don’t want my staff to be off site all day. But there’s other ways for people to give back and have this notion of social good. And one of the ways that we’re working on it, is we’re using our collective skill sets. And so we are marketers, we are data scientists, you know, we are advertisers, we are communicators. And we’re using our collective skill sets to do a deep dive into different data sets that will help us, Kate and shine a light on certain topics. And so one of those things that we’re working on is, we’re pulling together a couple of different data sets, to sort of tell the story, given that it’s now June and June is Pride Month. Chris, I know that you’ve been starting to experiment a little bit with some of the publicly available data sets. And really it, it all goes back to the core of what we do, we’re just applying it to a different context. And the core is sort of the six C’s of data quality. And the first thing is, you know, cleaning the data and preparing the data, making sure the data is usable. And so that’s sort of where you’re at in the process. So do you want to share a little bit about what’s going on, I do. But
first, talk a little bit more about choosing a cause based on the company values and things of it, because I think that’s an important part, a lot of people, a lot of I’ve seen a lot of brands, sort of do news jacking of, you know, popular trends or popular causes, with it being misaligned with what they stand for. And in some cases, y’all can really bite them if they’re doing something that is off brand.
Sure, you know, it’s it’s such a sensitive topic, but it’s such an important one where, you know, companies, when they take a stand on a political leaning one way or the other, they know they’re going to be alienating part of their customer base. And that’s just a risk that you have to take. And we are fortunate enough that, you know, the our entire company trust insights is for the most part aligned on all of the things all of the causes that we stand behind. And we’ve made it a point to say that anyone who has a problem with the causes that we stand behind, we probably don’t want to work with them anyway. And we’re okay taking that risk. And so some of those things, you know, that completely aligned with our values, especially with this being Pride Month is that inclusion that, you know, making sure that nobody feels like they’re sort of sitting off to the sidelines, and they can’t participate. And that’s something that we want to make sure that we’re highlighting. And so we’ve chosen this month to really focus on what that inclusion looks like. And, Chris, I know you’re working on looking at the hate crimes data set.
Yeah, well, so it’s funny that it there isn’t just one, there are multiple organizations like the Human Rights Campaign, like movement, advancement in progress of map, it’s called all these organizations that are trying to compile data about what’s happening with their, their audiences. And then there is the US government. So the data set I started with was the Federal Bureau of Investigations, Uniform Crime reporting, which is their central data system for a collection of all crime data. So this is everything from like your car jacking and all this stuff, but they have a whole special thing that was set up and set 2017 co author on hate crimes. And what they’re doing is they’re they’re getting hate crime data from police departments around the country. And they’re compiling it and publishing it. Now, the government’s no surprise, their data is not the cleanest. But it is very insightful, because they delineate six major categories of hate crimes such as race or ethnicity, gender identity, gender, itself, sexual orientation, religion, and so on and so forth. And they publish this by state. Each state has their own spreadsheet, they have to download all 50,
soda soda all together.
And you get a relatively small number of features to begin with to get the number of crimes in each category, the number of agencies in total number of agencies reporting, so that was one day is that that’s the starting date is that because that’s probably the most uniform, then you have to go to the US Department of Justice, the US Department of Justice, which is separate from the FBI. Well, I think the FBI as a part of it, has hate crimes themselves, what states have laws regarding hate crimes, and then have laws requiring the collection of data, because which is a separate table. And this one was very interesting, because just because something is against the law does not mean a police department needs to report on it. When I looked at the original data for from the FBI, there were a large numbers of hate crimes in like California and Massachusetts, and New York, and part of it is population based. But then when you look at the department justice table, you go, oh, there are a whole bunch of states where Yes, there are theoretically hate crime laws on the books, but there’s no law requiring police departments to collect the data. So like, for example, in Alabama, there is technically a hate crime law. statute. But there is no requirement the police departments collect the data. And then if you go back to the FBI data, and you see three out of 300, police departments reported this, like, oh, okay, so there’s that and then go to h RC and map. And they have data about which states have laws specifically about gender identity, and sexual orientation. And that was really surprising to me, because I think there are 23 states that have laws about protecting sexual orientation, like only 16 that protect gender identity. So there’s even more granularity. And so what we wanted to do with this project was how do we take all these different data sources, normalize them and sew them together? Because one of the things that got me thinking over the weekend was, there’ll be more people who would be interested in doing this kind of analysis, if the data was available in a format that they could process. If you tell somebody, hey, if you know, LGBT rights are really important. Here’s 72, tables, all which are different formats. Good luck finding insights, as opposed to here’s the one spreadsheet with all the features. See what you can see. There’s a big difference. There’s a big gap between those two. So that’s kind of the idea. Now, I love to hear your thoughts about what are the things that we should now start doing with this data set, start doing basic feature engineering, like what percentage of police department report comes? But what are the different ways that you would say, if someone wants to pick this data set up? How should they be thinking about it?
Well, I will resist the urge to start to unpack all the things that are wrong with, you know, the state laws and everything, because that’s a whole separate topic that is interesting, but probably not relevant for this particular podcast. So I’ll put that aside. But I do have very strong feelings about a lot of what you just said. But you know, I think at a very practical level, you’ve just described why more people don’t do this kind of analysis, just the fact that you have to get an individual’s data set for each state, and it may not have usable data in it is problem number one. And so, you know, when I think about the types of questions that we want to answer, and why you’ve decided to undertake this project, I mean, the first question I would want to know is, you know, what are those hate crimes consist of? And are there ways that we, as communicators, can do some education, so to help people protect themselves or to start to prevent some of those hate crimes from happening, because it’s a matter of ignorance? You know, and that’s a very lofty, idealistic view of the situation. You know, being a woman, I have experienced a certain extent of a version of hate crimes, but not nearly with the intensity that someone who identifies as LGBT Q. And so I think that that’s where we want to start to think about it’s, it’s one thing to just analyze the data and say, Oh, this is what it says, but what actions can we take? And so I think that a big call to action for us, is that education piece of, did you even know that this is happening right in your own backyard? What can you do to help prevent it? What can you do to educate? I think that for me, those are the some of the goals of this project.
It’s interesting, you say that, because in a lot of cases, I think the data is so so underreported, that it’s statistically not valid, I mean, not to draw stereotypes of again, get too political. But if you tell me that there were only nine hate crimes for the entirety of 2017, and the state of Alabama, I’m probably going to say that your data is a little bit off. And so the first thing that tells me that I would think of is
in terms of actions to take,
that I believe, I believe would be politically pleasing to different ends of special on the one hand, on the on the more liberal side of the spectrum, saying, Hey, we want to make sure that there’s fairness and equality. And on the other side of the conservative side, saying, Hey, we want to be tough on crime. I think there’s an argument to be made, let’s make sure there are laws that all 50 states require the collection of this data, right, no additional, you know, major initiative. Second thing, we want to make sure that this data is collected properly, and cleanly, so that we can then make that analysis because you can in the FBI, Did it break it down to the different types of crimes like violent assault, non violent assault, sexual assault, you know, property damage, but with nine cry, right, you have, you have an initiative. The other thing that you could do thinking about is from an education basis, if you know, the percentage of departments that report versus the percentage of departments that there are total in that state. If you would turn that to a ratio, you could, potentially and this is this is squishy math here, but it’d be a good starting point is to say, What if we adjust the number of cops by the underreporting ratio? So three out of 300? are reporting a 1% of reporting? Do you then add a new I’m 99% multiplier on to those nine crimes to say like, yeah, there’s probably more being more happening in the rest of the place. The other thing and this is important for, for people who want to do this kind of work, as well as all marketers is, make sure that you’re thinking about engineering those features, which is a fancy way of saying add relevant data from the data itself. So if you have three departments that are reporting, and 300, aren’t, you can create a ratio, if you know the number of hate crimes, and you have the state’s population. Now you have another ratio. From map, we got the percentage of census based people who identify as LGBT Q and in each state. So now you have a population of the LGBT Q, you add the other population of the state, you can start creating more of these ratios to help engineer a model saying, Are there is there there there about things like laws and data collection? That helps us understand the nature of hate crimes better? us a question for you.
How do you draw the lines? Or how do you?
How do you avoid letting your emotions influence what you’re doing when you’re working with sensitive important data? How do you you know, keep your in some way, keep your heart out of it so that you can be as objective as possible? What are the some of the things because as people start to dig into their favorite causes, of course, you’re going to feel emotional, you should you care about it, you don’t care about it, you’re probably not going to do very good work. But how do you do? How do you define that balance?
It’s difficult, and it’s not always going to be perfect. You know, one of the reasons why we are taking on this project is because we can so much about it. And so to your point, our hearts are already involved, you know, our emotions are already wrapped up in the data set. But I think, you know, through time and experience, you learn to separate the emotion from the logical, sort of those, you have those two different minds, you have your emotional mind, and you have your logical mind. And you know, in everyday use, you have to separate the two. And it’s no different when you’re working on a project like this, you have to set aside those emotions. And it’s easier said than done. And it’s something that if you’re newer to a process like this, then you need to build in those checkpoints, you know, every hour every day or so to really step back and see if Okay, am I still looking at this objectively? Or am I getting too emotionally involved in this. And that’s why it’s always good to make sure that you’re doing peer reviews, or more working with a team so other people can gut check the work that you’re doing. It’s not easy, especially when you’re working on sensitive material like this, you know what you were saying, Chris, about sort of those actions it was it was making me think that when you think about the data analytics hierarchy, what you’re doing is really pulling the what and so that’s the bottom rung and the information, the insights that you’re going to get from this. Quantitative data is going to allow you to start to dig into the why that qualitative data, so it will give you questions that you can ask in market research surveys, or very targeted surveys state by state, to really ask those questions of Did you know that there are laws? Did you know that this was happening? Do you know what a hate crime even is? And so you can really start to unpack each of those different pieces and start to get to the why, like, why is it not being reported, we have assumptions as to why it’s a very difficult thing to do. But actually having that concrete data from, you know, a representative sample will help, you know, just add more depth to that quantitative data that you’re pulling.
I completely agree with you the caveat,
as as you alluded to, is that the data collection itself can be very, very tricky. When we are talking to anyone who’s been the victim of any crime, there obviously, are medical issues. I mean, you’re you’re basically crossing over into HIPAA, you’re crossing into that late, protected, personally identifiable information you’re talking about, you’ve crossed over to protected health information in many cases, because if you are the victim of assault, particularly sexual assault, that is, by definition, a medical thing. And one of the things that was in some of the FBI data in terms of the caveats that came along with that is it said, you know, there are, there are reasons for under reporting, a lot of which can be feeling like your information stopping sector that you are not anonymous, that you may be, you may face further retribution. And also, one of the other things that is interesting in in the FBI is data is that it says that the crime, if it to be classified as hate crime must have been done with a stated bias. So somebody, you know, for example, I’ll use a very silly example, someone who attacks a pasta, Ferrari and with a plate of bacon as opposed to a tomato sauce. If they don’t state that their motivation is that they hate pasta, Aryans, then technically, at least, the FBI is definition may not be classified as a hate crime, where as we know, from practical, everyday experience, if you go and scroll a swastika on a Jewish temple, that’s pretty clearly a crime. But may not necessarily there may be differences, like you said, in the subjectivity of the reporting. So part of the action that we can take is to encourage a uniform standard of reporting like this should be classified as a hate crime, this should not. But even then that that gets tricky and messy, because now you’re talking about a different type of personally identifiable information that of the attacker. And that is also information that you still need to treat with sensitivity, because our system of law is such that you are innocent until proven guilty in a court of law.
Right. And I think when I’m talking about doing those qualitative surveys, I’m not talking about asking the people who have reported it, it’s more of that broad strokes of, you know, sort of getting that pulse check state by state, the awareness that people even have of the situation and the laws and the categories and what these definitions mean. And you know, we’re talking about it in such a way you and you and I are working on this independent of any organization, but one of the ways that people can get into involved is to partner with one of those organizations that really carries the torch for that cause. And so whether it’s, you know, the human rights organization or another organization that’s, you know, really fighting for that equality for LGBT Q, you can partner with them, you can offer up some of these ideas that you have of, hey, I want to do this thing. Are you already working on some of this, you know, and so can I help you, or, here’s some things that I want to be able to do for you to help you because a lot of these are nonprofits, their fundraising, their volunteer based, and so if this is a cause that you feel passionate about, go ahead and start volunteering with them. And volunteering doesn’t always mean, you know, standing outside or collecting money, it can be a lot of what Chris and I have been talking about using the skills that you already have to help these causes, you know, move forward,
it can be one person sitting in their basement on a Saturday night. But just spreadsheets,
it actually absolutely can. And I think that that’s such an important thing to note. Because the way that you can give back as a person to the causes you care about it, it can always look different. And so Chris, the work that you’re doing, it’s doing a couple of things. One is it’s keeping your personal skill set fresh. And it’s really challenging you outside of some of the project work that we currently have at trust insights. But it’s also giving you an opportunity to dig into a cause that you care about and give back to that cause without doing any sort of monetary fundraising or actually volunteering your time, at you know, a different location.
So to wrap up, if you are if you’re interested in supporting a cause, we would encourage you to go out talk to a cause, find a cause you believe in, see what they’ve got, what they’ve published already. And then what are the ways that you could add your own analytical or experience based insights to that dealt amplify it to help make it better, to publish a new spin on it in some way that could be beneficial to it. But no matter what we would encourage you to do go and do something that would be beneficial to a cause you care about. As always, you can find us at trust insights.ai have questions about your marketing data and analytics, please subscribe to the YouTube channel and to the newsletter over at trust insights, AI and thanks for listening.
And don’t forget, it doesn’t cost anything to be a nice person.
Thanks for listening to any your insights, leave a comment on the accompanying blog post if you follow up questions or email us at marketing at trust the insights.ai If you enjoyed this episode, please leave a review on your favorite podcast service like iTunes, Google podcasts, Stitcher, Spotify. And as always, if you need help with your data and analytics, visit trusted insights.ai for more
Newsletter
Podcast
