#039: In Defense of Excel

The following is a straight-up machine translation. It has not been human-reviewed or human-corrected. However, we did replace the original transcription, produced in 2017, with an updated one produced using OpenAI’s WhisperX in 2025, which, trust us, is much, much better than the original. Still, we apologize on behalf of the machines for any text that winds up being incorrect, nonsensical, or offensive. We have asked the machine to do better, but it simply responds with, “I’m sorry, Dave. I’m afraid I can’t do that.”

00:00:04.00 [Announcer]: Welcome to the Digital Analytics Power Hour. Tim, Michael and the occasional guests discussing digital analytics issues of the day. Find them on Facebook at facebook.com forward slash analytics hour. And now, the Digital Analytics Power Hour.

00:00:24.73 [Michael Helbling]: Hi everyone, welcome to the Digital Analytics Power Hour. This is Episode 39. Talk to any good data scientist and eventually you’ll get their opinion on Microsoft Excel. It’s usually negative. And does Excel deserve those slings and arrows? Or is Clippy just as useful today as he was in the past? I can’t even get through the intro. Tim’s already laughing. But no, that is the question we’re going to talk about. What is the place for Excel in the world or the toolbox of the digital analytics professional? As Clifty might ask, it looks like you’re attempting to listen to a podcast. Would you like some help with that? Because here is Tim Wilson, senior partner at Analytics Domestified and quite frankly one of those Excel gurus that you’ve been hearing about.

00:01:15.17 [Tim Wilson]: Actually the senior Excel correspondent at Analytics Domestified.

00:01:21.64 [Michael Helbling]: And I am Michael Helbling and once upon a time I taught a class on Excel to some people. Alright Tim. Should we just uninstall Excel? Are we ready for that as an industry? Or does Excel still have a place? And if so, what is that place?

00:01:37.85 [Tim Wilson]: It is hard for me to imagine a world where there is no Excel. We’re hitting this topic in a point where I am highly conflicted, I think for one reason. I think you may be conflicted similarly. I’ve been a long time user of Excel. I have taught classes many times on Excel, a big proponent of it, but at the same time, I feel like I’m hitting a point where I see other tools that really can go beyond Excel. I’ll say when Tableau was a little bit more in its nasancy, and a lot of times people would say, oh, Tableau is so much more awesome than Excel. And then more often than not, I would see what they would do. And actually, this is still the case. There are people who just rave about Tableau, and their output from it is, I guess, kind of underwhelming. And we can kind of debate. I don’t know that we can debate, but there are times where they say yes, but it was so much faster in Tableau. And then I come back and say yes, but it’s so much less portable. send that Tableau file to the 99% of the business users who won’t be able to see it outside of a PDF and I will send my Excel file to every one of those who will be able to open it. So I think it has its place absolutely. I don’t see it going anywhere soon, but at the same time I’m finding in as It has been mentioned many times on this before as I’m spending the year exploring R. There are lots of things I’m finding in R that I’m saying, wow, I couldn’t, I wouldn’t even attempt this in Excel. It would be such a nightmare, even as a fairly sophisticated Excel user.

00:03:11.59 [Michael Helbling]: What do you think? Well, no, I’m in the same boat. I think Excel gets used by default by so many businesses. In fact, it always surprises me, or maybe it doesn’t surprise me anymore, but it used to surprise me a great deal. How many companies, really big ones, had so much of their operational capability tied up into some sort of business running Excel workbook. And if that thing ever cratered and disappeared, like there would be big trouble and a lot of important data and important data manipulation was happening. And a lot of times what I feel like the main complaint is that people do tend to sort of push out past the boundary of what Excel maybe should be doing and attempt to do too much. because Excel gives you avenues to go down. It’s got stuff in it to do statistics, and it’s got stuff in it to do data visualizations, and it’s got stuff in it to, you know, if you put some things in there, like, Power Pivot or Power BI or whatever a lot of really powerful tools inside of it But it’s you start going out past what it’s really good at and certainly more recent versions of Excel have expanded its capabilities even more You know I used to work in a world where you know We had the we had the luxury of being blocked at what was it like how many rows was it?

00:04:37.57 [Tim Wilson]: Was it 10,000? It went up to a million.

00:04:40.15 [Michael Helbling]: It jumped to a million.

00:04:41.16 [Tim Wilson]: It was 64,000. Was it just 64?

00:04:45.63 [Michael Helbling]: I thought it was in the hundreds of thousands.

00:04:47.57 [Tim Wilson]: No, no, no. It was two to the whatever, the 64, five, six, or something.

00:04:53.40 [Michael Helbling]: It definitely jumped to a million. It was a huge jump when it went to a million. That tells you I’m not an Excel guru because I don’t remember that number off the top of my head. But it used to be back in the days when we would try to pull a log file into Excel because, you know, why not? You can only get so many rows.

00:05:12.55 [Tim Wilson]: Well, I remember when they jumped to a million and I actually took a log file and threw it in and I said, well, is this going to, is this going to crater it? That it’s a theoretical, it will allow me to have those rows, but in practice it’s not going to let me. And it actually handled it pretty well. I was kind of heading a different way or thinking a different way when for the companies that are, I’ve definitely seen the spreadsheets that are these monsters and there’s, you know, like one guy. who knows how to update it. And it’s kind of expanded and bloated over time. Moere often than not though, I tend to see that as the fragility tends to be because it wasn’t really built out with a plan. And I’ve, you know, at unsummit and the class I did a couple of months ago, like when I talk about like building dashboards, part of the core thing I talk about is having like a good structure and so that you can kind of extend it and it’s scalable. And I’ve seen a lot of those spreadsheets where it’s this massive like bloated thing and then you look at it and you’re like, but it’s also impossible to follow. There’s no organization. This is like, it’s the equivalent of looking at somebody who just wrote shitty, shitty code and somebody saying, oh, this programming language is terrible because this thing is, you know, fragile and bloated. And it’s like, no, you, you, you built it wrong. Well, and, you know, VB script is pretty terrible. Uh, B.B.A.?

00:06:33.96 [Michael Helbling]: Yeah, that whatever.

00:06:35.80 [Tim Wilson]: Well, that’s maybe a good 65,536 if my math is correct is the number of rows.

00:06:41.81 [Michael Helbling]: That’s exactly what I was thinking.

00:06:43.66 [Tim Wilson]: Sitting with my HP12C. What’s going to happen is when this is being edited, the editor is going to Google it and say, that was dead wrong. Do we let Tim spew off? Oh yeah, we’re totally going to.

00:06:55.52 [Michael Helbling]: No, well, and that’s that’s I think what it is, is that Excel is really useful for some things. And honestly, dashboards, I feel like using Excel for dashboarding is actually potentially great use of Excel, depending on what’s in your dashboard and what you need to show.

00:07:12.89 [Tim Wilson]: Well, and that’s big. I think maybe part of that is Excel, whatever the version, wherever they started making it easier to pull in data from external sources. So I was with report builder for Adobe or for any of the shuffle pointer super metrics, or you name it for. GA, you can pull it in, but then also even linking to an HTML to an XML, you know, URL and pulling the data in. So you can pull in and refresh the data. And I agree. I think dashboards because of their portability and because I built a couple of weeks ago, just a simple little mini dashboard and I had put on the part of what they wanted was like the top 50 of X and I built this nifty little thing and I had it going as a PDF, because it was just this nice little two page thing. And one of the requests came back and was, hey, can we get that in Excel? Because we want to be able to copy and paste those top 50 out of there. So I’m like, oh, I guess the report, but the simple report build a query on tab two is really what you want. But when people want to be able to take data and put it somewhere else, you know, there’s a downside to PDFs, definitely. But I thought you were going to be Is I am leaning towards R and I feel like you have been doing exploration around Domo. Is that seems like kind of another Excel alternative, right? That’s kind of part of their pitches. Get out of Excel, get into a more powerful environment.

00:08:33.36 [Michael Helbling]: Well, and I think for businesses that are running like big chunks of their business in Excel, getting out of that and sort of not being stuck with just one person who knows how this whole thing works. And if they get hit by a bus, our whole business goes with it. And that’s where stuff like DOMO can come in really handy. But again, like with any technology, I always look at it on a continuum in terms of your sophistication, your capability, whatever it is. And so as you kind of progress towards being able to maybe utilize this data more collaboratively across teams or something like that, then tools and systems like DOMO or sweet spot become Way more compelling one of the reasons why I mean like one of the things that you’ll hear like the from the sweet spot guys is just standardizing across Sweet spot intelligence just to be clear.

00:09:27.92 [Tim Wilson]: You’re not talking about the sweet spot and you’re talking about the sweet spot intelligence

00:09:32.21 [Michael Helbling]: Sergio Maldonado world. Exactly. That’s right. You know, standardizing across a lot of different groups is something that’s difficult to do with Excel because too many people can kind of manipulate all the things that are in there unless you really kind of lock down various pieces of it and so on and so forth. And so a more single-purpose tool like Sweet Spot or DOMO Would be more useful for kind of curating those aspects of your data. So Yeah, you you I think there’s multiple boundaries, right? There’s the boundary of we talked about first like just data limits, right? How do I what’s the point or the boundary from when I go from Excel into something that’s better exploring bigger pieces of data? Then there’s the analysis side, which you were talking about with R, like when’s the, what’s the frontier from Excel into deeper statistical tools? And I think what’s the boundary from Excel into better visualization or communication tools? And so look, those are the kind of three main categories. I think I just gave three main categories of where, if I said four, that’s a bonus category. of where you work with Excel up to you get to that point of, okay, now I need to go and do something else. I think partially too, there didn’t used to be options outside of Excel for a lot of companies. In the last 10 years, we’ve seen a lot more options out there for people. And now somebody’s gonna come tell me I’m super wrong about that.

00:11:00.83 [Tim Wilson]: Well, although you think back to, as you were talking about some of the control and the, so I worked at a place where certainly we used Excel a lot, but we had Cognos and we had PowerPlay. So we were building cubes and it was the BI team was heavily involved in like what are the cubes and then lots of people could add access to the cubes. And that was kind of an environment where they could safely slice anything within that environment and the cubes were being sourced from the warehouse. I still feel like a lot of times they’d get to something and they would copy it and paste it into Excel. But when you talk about those three or four, however many buckets, it is one where Excel has been around long enough and as easy it is to beat up on Microsoft, they have steadily improved the maturity to do all of those, right? Their data visualization has steadily gotten better. In some cases, they’ve added functionality they shouldn’t be giving to people, you know, 3D cone graphs. Nothing tells a story like a 3D cone graph. You know, but from a, you know, storing the data, like to be a little, a, a, a mart, right? When they go up to a million rows, which I think was at least 10 years ago at this point, when it comes to allowing the manipulation, you know, how can I share, even putting the statistical functions in, you know, the analysis tool pack, they, they sort of wisely said, it’s, you got it, it’s free. You just have to know you have to be smart enough. to go at it before you, you know, start clicking on stuff in your head explodes. And I think it’s got kind of a lousy experience, just like the query builder tool, which still looks like it’s from like 1997. So yeah, it’s one of those where it does the it’s the broadest, but maybe not, not all that deep. But what do you say to people and maybe it is the data scientists as you set it up who say, ah, you know, that tools, yeah, and nothing.

00:12:53.17 [Michael Helbling]: Well, and in fairness, like I try to look at it from both sides and I definitely come from the business user business analyst side. I’m not a data scientist. So when I’ve heard people say that, I thought, okay, what is it from their perspective that makes Excel so not a great tool to use? And I do think it has to do with flexibility and options. A data scientist is thinking and working very quickly through a number of different things. They need to be able to just very quickly take the data any direction they see it going without worrying about those boundaries. The other thing I think is that Excel has been so entrenched that probably really smart data scientists have seen companies, people, business units, whatever do just really dumb stuff. just so they could do it in Excel, as opposed to leveraging a much, much better tool for that purpose, for whatever that reason might be. And I think that probably builds a sense of frustration of why can’t we just use good tools or appropriate tools for the job that you’re trying to do. And I’m trying to think about the last conversations I’ve had with people in the data science side of things and sort of what I remember hearing them say, but I feel like that gets to probably where some of that frustration lies. It’s like, I would never, I’ve heard people say, I would never willingly use Excel for anything.

00:14:17.32 [Tim Wilson]: And it’s almost- You’ve heard those are the data scientists type people? Yeah.

00:14:20.58 [Michael Helbling]: Yeah. Okay. And because like the things that they’re doing, they can, they have better ways of doing all those things with different tools. But then, then, but then my argument back is, but yeah, not everybody’s going to have the access and intelligence and training, training to work with all of those tools. I can barely get my head around a left outer join or something, right? So. So. you know, then for me to be like, you know, conceptually hadooping it up, that is not going to happen probably.

00:14:54.67 [Tim Wilson]: And that’s maybe that is from a iterating on the data. I think of some of the other tools that I’ve kind of crossed paths with, you know, Clementine from SPSS. I don’t know if they bought it from somewhere. It was kind of a visual laying out a sequence of doing things with data. and that’s what sort of had a on the fly ETL but then also kind of you know manipulation of the data and that is one of those if you’re saying I start with the data at in situation a and I want to join it with set B and I want to do this and do this and then I want to rapidly iterate through doing a bunch of stuff you know once you’re in the analysis tool pack and using the solver It does feel like all of a sudden you’re, you’re working with a dialogue box where you’ve got like eight things you can set. Now 99 and a half percent of the people using Excel don’t know what those eight fields mean. You know, the residuals, what are you talking about? Which I don’t think is solver. That’s regression, maybe. I don’t know. Just ordered a new book on statistics. I’m going to take my third and final run at it and then quit and go paint trees on PBS. Ooh, there’s a pretty cloud. But can you think of an example or even a type of example when you say people have seen companies do things that don’t make sense or that are a bad idea just because they could do it in Excel?

00:16:16.32 [Michael Helbling]: Well, anytime you see companies kind of pushing Excel out past sort of its natural boundaries, you know, whether that be Trying to use it as a database trying to extend it out past kind of like where it should be from a data sharing or data collaboration perspective using it. I mean anytime you’re using it to like manually gather multiple sources of data.

00:16:40.29 [Tim Wilson]: Using it as your marketing calendar that everybody’s going to update the Master Bulleted Excel file.

00:16:46.16 [Michael Helbling]: That’s one of my least favorite. I mean, we get into Google Sheets at Search Discovery and if we’re all collaborating at the same time, it turns into a mess. Because we’re usually, it devolves very quickly into like people actually entering fields to trying to make fun of whoever else is in there and or add in animated GIFs or something like that.

00:17:10.26 [Tim Wilson]: which Google Sheets is another interesting one that does feel like it is carved off. If Excel is trying to cover a very wide span, and I do think there’s still, I still will contend that there are times when people say you’re pushing Excel too hard and it’s like, no, you’re just pushing it the wrong way. To me, it’s become a very, very stable platform. You know, there was a time when it would just, you know, your file would spontaneously be corrupted. And once it was corrupted, you’re fucked. But as in sheets started out as being this super simple, like, oh, remember, you know, Lotus one, two, three, you know, Excel 1995, Excel 95. And it was like, it’s basically your rows and columns, and you can put formulas in it. And because it had the collaborative ability, the it’s in the cloud, it’s you’re not going to lose it because your hard drive crashes. It sort of, you know, has started pulling from that lower end use of Excel. And there are many, many cases now where if I don’t have these heavy visualization needs, hell yeah, you know, use sheets. So that’s actually carved off to me a big chunk of what was, and has made a vast improvement to what Excel can do. And I don’t know, Excel has their kind of cloud based solution, right? All the stuff they’ve been doing with SharePoint integrations and I don’t really even Do I still operate with the individual files? Um, but it seems like they’ve kind of tried to, to head that off a little bit to be a little more collaborative and cloudy.

00:18:33.34 [Michael Helbling]: Yeah. Oh yeah. Well, and the latest version of Excel is actually going to support or is supporting JavaScript. So I think for, for certain things and data transformation and things like that at JavaScript, it’ll be a much more accessible language or data automation.

00:18:49.20 [Tim Wilson]: So I’m going to, I’m going to find out that I’m a better BBA programmer than JavaScript writer. Oh, dear Lord.

00:18:55.28 [Michael Helbling]: Yeah. Well, I mean, you know, I have people who helped me with that. I’m not good at either one. So let’s talk a little bit about some of the boundaries. So if someone’s listening, okay, let’s, let’s talk about data visualization and using that to communicate a story around data. We talked a little bit about dashboards. That’s one method of communicating data, but there’s also analysis. There’s reporting. Where do you see the boundary between, yeah, that’s good. You can keep using Excel for that. Or, all right, let’s talk about other tools now. We need to put you in a better spot so you’re not kind of doing more harm than good.

00:19:36.93 [Tim Wilson]: So I like that way to look at it. So with visualization, so with dashboards, because they’re inherently, or even for reports that require some level of formatting, that are going to live on and be updated, refreshed on a daily, weekly, monthly basis. You can afford to put a lot of care into exactly how should this data be represented and is it going to continue to work even if the data fluctuates. When it comes to I’m presenting this data one time, I’ve done an analysis, I found my thing and let’s say that I found it in Excel, I was crunching it and now I need to get a visual that I want to show, you know, ultimately it’s going to go in a presentation, PowerPoint, Keynote, whatever. I feel like there’s, it’s almost like half of the time. Excel is perfect. If this is going to be a horizontal bar chart, it’s going to show what I want. I can accent my one bar that’s highlighting the thing. I drop it into my presentation. Maybe I draw an arrow or put an annotation next to it. Or as Lea Pica has slowly gotten me to realize that if I’m doing that, a lot of times it makes more sense to use the native chart builder in PowerPoint because that’s basically Excel under the hood and it’s not as clunky as it used to be. Where I’m starting to waver a little bit is when it comes to doing a Venn diagram in Excel, you’re kind of out of luck. You can calculate how big things have to be, but then you wind up really hacking stuff, you know. And even I’ve done, I feel like I’ve done sort of crazy things with bubble charts where I’m just kind of forcing circles to move around to arrange them how I want them just so I can copy and paste it. I have a sense that that’s the sort of thing that Tableau does way better. And then even as I’ve been diving into R, those are some things that I feel like, oh, with the R has a completely different paradigm for charting or the ggplot2 package where it’s like forget everything you know about the basics of a chart and the chart types there’s really only one chart type and the way that you describe it and define it you can make anything and I haven’t fully internalized that language yet but it does get to you know much I think faster richer opportunities to say You’re not trying to make something a visualization nobody’s ever seen before. You still got to be using the language of lines and dots and circles and bars. But I think six months ago, I would have said, that’s bullshit. I can visualize anything I want to in Excel. And I think I’m changing on that. And maybe part of that’s from the number of data visualization blogs and conversation going on. People point to things where I’m like, I’ve never seen something visualize that way. And then I think, and that’s a really effective way to visualize it. And then I think I could never do that in Excel.

00:22:15.40 [Michael Helbling]: Yeah, so it comes down to as your toolset expands and as your visualization capability expands, you start needing or wanting more ways than what Excel gives you to present data.

00:22:29.35 [Tim Wilson]: Yeah, and Excel, you know, one thing Excel doesn’t do animation, really. I mean, it can.

00:22:34.56 [Michael Helbling]: I mean, chandu.org has all sorts of examples, but yeah, I remember downloading some juice analytics ones where you could use the offset function to like tab through and update your You know, it was cool little trickery. And that’s actually the thing is creative people can take and stretch Excel in so many cool different ways. And have done, right, because necessity was the mother of invention, have done some amazing things inside the confines of Excel. And that’s also maybe where people get frustrated because it’s like, why apply all that creativity here when you could apply it over here and get faster, better results?

00:23:10.70 [Tim Wilson]: And even if you, being one of the people who has hacked it a lot, waterfall charts or another one, can they’re kind of, they’re kind of famously excel, like you got to go and hack all sorts of bar chart shit with no fills. I can do it, but, and I can do it, you know, I can, I can say all day long, you know what? Cause I know what I’m trying to do. I can do it pretty quickly and pretty efficiently, but come on, like I’m, inherently hacking something and experimenting and tweaking. And gee, if I had somewhere else that I could say, I’m doing a waterfall chart. So let’s go with it.

00:23:42.58 [Michael Helbling]: Yeah. Like the first few times you tried to make a bullet chart or something like that and Excel and it’s just been like going through all this pain.

00:23:52.51 [Tim Wilson]: Yeah. But yeah, yeah. Cause there’ve been blog posts, how to make a bullet chart in Excel and exactly.

00:23:57.22 [Michael Helbling]: You can do it.

00:23:58.54 [Tim Wilson]: Step one, but at the same time along the AC line. if you’re doing that, if you’re doing a dashboard where you have 10 bullet charts, you know, it is one that if you do it right, once you have it, you know, okay, you’ve put that extra hour investment in once and that’s nothing, you know, six months in or three months into it going out. But yeah, and you’ve made it portable. And I would say there, this is probably a goofy little benefit of you know, as analysts, we want to be included in the conversations because that’s what we need to be. And it has happened many times. I’ve gotten roped into doing Excel classes because that’s what people really wanted to learn. They wanted to be more efficient because they would, they would be doing something incredibly tedious. They’d happen to ask me and I’d have it done in like two minutes. And they’re like, Oh my God, like that’s very helpful. And so it.

00:24:49.56 [Michael Helbling]: Yes, edit, fill down. And they’re like, what?

00:24:54.71 [Tim Wilson]: Well, so it is a way for the analyst to get kind of this is the analyst has telling that the analyst can help you clarify your objectives and establish your KPIs and help you with, you know, validate hypotheses is all kind of squishy. If it’s, hey, the analyst can help you not spend the next the entire day, you know, manually doing something annoying. And so there’s a way to get kind of in good graces and have an instant kind of solid relationship and really helping a business user, not necessarily running the business but giving them more hours in their day and they’re going to remember that and then they’re going to remember to come ask you the next time and the next time and as long as they don’t think that’s all you do, you’re somebody can help them out and there’s value in that. If I tell them that Yeah, they need to tell me exactly what data I want so that I can set it up and pull it with R or Tableau and then I’ll send them a PDF and it’s got to go into my queue. That’s not necessarily as helpful.

00:25:53.79 [Michael Helbling]: All right. So let’s jump over into the analysis side of where’s the frontier. Can you use Excel for analysis or does it, you know, is your cliff, where’s your cliff, where people need to start thinking other tools? Here I’ll I’ll start this one because your answer will be way better than mine. So I got to get my licks in early. I’m starting to believe and this might, you may be having an impact on me, but I’d like to think that I’m just really intelligent is I’m starting to believe actually.

00:26:27.34 [Tim Wilson]: And therefore you’re moving forward despite me having an impact on you.

00:26:31.45 [Michael Helbling]: Well, I’m just saying like, you know, the fact that you and I might agree on something along these lines should not be viewed as anything more than purely coincidental. There’s no significance to this. That’s right. Correlation is not causation. So I’m starting to believe more and more that real analysis isn’t something that you do very much of in Excel. So when I simplify analysis down to its basic fundamental thing, it’s learning the difference between two things. And more and more good analysis encapsulates the data at its lowest level, not the data aggregated and exported from reporting tools like Adobe Analytics and Google Analytics and so on. And so the rise of tools that allow us to export the lowest level data, the event level data, the hit level data. and then do analysis on those data sets. And if you look at where BigQuery fits into Analytics 360, we look at where Analysis Workspace fits into Adobe Analytics. This is the usage of some of that hit-level data to create better analysis capabilities. Excel, I think, quickly runs out of room to be the right tool for analysis. I don’t want to say R is a better tool, but in essence it becomes a really good tool in the tool set because it can do good analysis at that point for that type of data or other types of analysis tools as well.

00:28:05.23 [Tim Wilson]: Well, so yeah, I mean, we wind up in the definition of what is analysis. And if I say if we if we agree that analysis is validating hypothesis, it kind of depends on, yeah, it depends on the specification of two things.

00:28:18.20 [Michael Helbling]: It’s this is different than this. And I can prove it with by these methods.

00:28:22.28 [Tim Wilson]: So I feel like I have so using Excel as using it smartly and giving this gives it gets a little to the discussion of giving business users enough data that they can play around with it without hurting themselves, you know, exporting. a data set where you’ve got one way to get more granular data is to add more dimensions. So if you have a bunch of dimensions and you have a bunch of metrics and let’s not talk about users or unique visitors, then you instantly you quickly get a larger data set more rows in your table. And then you’ve got the ability with a pivot table to look at that a whole bunch of different ways and even with calculated fields. So I do think that there are cases where enabling marketers to do self-service analysis by thinking through and building out that kind of master flat table, giving them pivot table, giving them slicers, giving them kind of visualizations tied to that. That does enable some level of exploration that I’m sure the data scientists would say,

00:29:25.25 [Michael Helbling]: Analysis because that’s not applying any sort of statistical Yeah, it’s just doing a sort of a comparison but those okay, so I do love me a good pivot table So that you that one took me back a step from where I started so that’s good

00:29:41.55 [Tim Wilson]: So but but at the same time if I say well gee now I’ve got that big flat thing and now I want to take a one more level either I want to take a one more level to super granularity or if I don’t want to rely on what I decide to slice and filter by I want who to have this is not machine learning but this is I want a machine to I know what I’m looking for I’m looking for outliers when I look at this metric crank through all the different ways you can do this. To me, that’s one example of where no Excel doesn’t do that real well. I think after our last, I went back to, I do have John Foreman’s DataSmart book. I realized that it’s not a good book to have as an ebook because you’re supposed to be doing exercises. So it’s not good bedtime reading. But he talks about Solver as one of those things, but I still feel like Excel with tools like Solver is kind of constraining, you’re being given this very powerful tool to work in a very narrow situation. You have to recognize it and then use it. And I feel like some of the other tools are a little more broad, where you’ve got a little more control over the number of inputs, the number of outputs, and you can kind of configure it more, but I am absolutely schizophrenic on this entire subject right now.

00:30:56.43 [Michael Helbling]: Well, but I think it’s okay because actually a couple of really good things I think popped out at me anyways and hopefully that’ll mean for a couple other people too. You know, if you are a data scientist and you’ve made it to this or if you’re one of the two people listening to this podcast. Yeah. Okay. So then the third one was around, I guess data manipulation. So I think you just hit the boundaries of that when you run out of memory. I mean,

00:31:20.36 [Tim Wilson]: Well, no, but I think you also hit it from a data manipulation, there’s a traceability part to it. If you bring it into Excel and say, I’m going to remove the outliers, I’m going to sort by this, I’m going to do this. As soon as you start manipulating the data, it’s much harder unless you’ve written a macro to do it to have traceability. And I think that’s these other tools that follow more of an ETL.

00:31:42.62 [Michael Helbling]: And I feel like that also is part of what you’re doing as part of analysis too. So I feel like maybe it’s not so easy to break those two apart. But so some of the similar points I think still apply.

00:31:54.73 [Tim Wilson]: But I mean, I think that I’m so spinning like every free moment that I have like fucking around with R the This is kind of helping me. I’ll clarify some of this stuff that repeatability. I remember Tom Miller telling me this at Emetrics Boston, I think. He was like, it’s repeatable stuff. And the way that you’re working on it, you can do it where you’ve got no audit trail, but you pretty quickly learn that the way you should be doing this is in a way that you know you can go back and repeat. You’re keeping track of all the steps that you followed. And Excel is not really great at that. You kind of need to be diligently taking notes of what you did with it and where you got the data. Right.

00:32:33.96 [Michael Helbling]: I remember Eric Goldsmith on our episode had some similar points. And that is pretty important stuff. All right, well, let’s pivot this. Let’s table this discussion and pivot to our last call. This is a portion of the show where Tim and Michael talk about things that they’re interested in that have happened lately in the world of analytics.

00:32:58.87 [Tim Wilson]: So I’m going first.

00:33:00.30 [Michael Helbling]: Yeah, go for it.

00:33:01.04 [Tim Wilson]: So I got a, I’m going to give you the choice. I can go with the, I’ve got a, I’ve got a very tactical tool, free tool thing, but I don’t want to wind up in a rut where all I’m trying to do is come up with a tactical tool last time. I did. So there’s that or there’s the completely whimsical that is a stretch to say how close it is to analytics, but it’s got, it’s got some geekery in it.

00:33:23.08 [Michael Helbling]: Let’s do that one.

00:33:24.13 [Tim Wilson]: Okay. So let me couch this by saying I in my life have read one graphic novel. So I am not a graphic novel guy, not even really a comics guy, but, and I’ve also think only wound up as a funder of one Kickstarter project. and those two come together in a graphic novel called Geeks and Greeks. And it is a very entertaining, light read. The guy who wrote it was, I’ve sort of crossed paths with him. He was about like five years older than me. He was in my fraternity at college, but he like graduated, I think literally the year that I was going to show up. So I don’t think we crossed paths. He’s had a, his name’s Steve Altus, and he has been, he’s been kind of a Renaissance man. He was actually in Die Hard with a Vengeance as one of the German terrorists. So he has like a… For a claim to fave. Yeah, he has a BS in aeronautics and astronautics from MIT, as well as a couple of master’s degrees. And then he wound up as an actor. And I think he’s been like motivational speaker and writer. And so he had this idea for a graphic novel about the Greek system, the fraternity system meets the nerd system loosely based on real life events at MIT when he was an undergrad and has this graphic novel. And it’s If you search for Geeks and Greeks, it’s kind of a crappy little site that you can buy it. But I got as part of my Kickstarter, I got the ebook version. And it was fun enough that I ordered the print version. A guy named Andy Fish. And I think maybe his wife maybe illustrated it. But it’s fun. And it’s a little bit of a gag. So it’s a nerd doing nerd things in a fraternity culture. So there you go. I think you’ve said enough.

00:35:18.39 [Michael Helbling]: That’s high on the whimsical. No, that’s awesome. We want the last call to be whatever is interesting to you right now. I recently finished it. And you, however, can go the other extreme. Well, recently, not too very long ago, everyone’s probably heard about it now, but Google announced Firebase, which is actually a way to build applications, mobile applications, but in the middle of all that is an analytics platform, a brand new analytics platform, and that is very fascinating. to me. And I’m excited to see the ways that it will sort of overtake or be part of the mix in the mobile application analytics space. You know, we’re doing more and more of that. And so I’m getting more involved and more excited about the potential for those kinds of things. So yeah, I’m excited to see about it. I’m sure by the time this episode airs, there’ll probably be some amazing stuff that have already happened because it’s a couple months since the announcement now. All right. So that’s my last call. All right. Well, Tim, I believe we saved Excel today. So great job. But we also found out when not to use it.

00:36:30.26 [Tim Wilson]: I think we found that it’s not black and white and anybody who wants to treat it that way.

00:36:35.83 [Michael Helbling]: We would love to hear from you as well. I imagine that the rate of inaccuracies, at least from my part of the podcast, has been very high. And Tim would like nothing more for all of our listeners to mention them on our Facebook page or on the Measure Slack.

00:36:53.42 [Tim Wilson]: Or hell, even go to iTunes and write us a review.

00:36:56.70 [Michael Helbling]: Well, maybe not after this show. Until we get a really cool guest to come on, then go right through view on ice. But as always, it’s a pleasure. Tim, thanks for taking the time. We’d love to hear from you, so drop us the line. And this is Michael Helbling saying, keep analyzing in Excel.

00:37:20.45 [Announcer]: Thanks for listening and don’t forget to join the conversation on Facebook, Twitter or Measure Slack Group. We welcome your comments and questions.

00:37:28.72 [Announcer]: Facebook.com forward slash analytics hour or at analytics hour on Twitter.

00:37:44.53 [Michael Helbling]: Bring me a go. Yeah, I’m with a lot of energy.

00:37:48.56 [Tim Wilson]: Okay, take it away then, Koss. What would you like Senior XL correspondent for Analytics to Miss Define?

00:37:57.66 [Michael Helbling]: Like a little daily show action?

00:38:08.24 [Michael Helbling]: Have you ever seen the odd couple, the original movie? And uh… Oh, there’s Jack Lemon. Jack Lemon. So he used to make these noises.

00:38:17.41 [Michael Helbling]: Oh yeah, that’s funny. Apparently people follow me on Twitter. Okay then.

00:38:29.13 [Michael Helbling]: Oh wait, I’m editing, so we don’t have to make that edit, I think. Yeah, if you’ve made it this far, probably just go ahead and shut it off, or it’s not gonna get better. No, I’m just kidding. Maybe they are on both sides, but it’s still independent. Because they’re Americans. Yes, what it is. Save a horse ride a cowboy. And this is Save a… What? Save a… It’s a name of a song.

00:39:07.75 [Tim Wilson]: Data scientists should get over themselves and not spit on everybody who uses Excel.

00:39:11.53 [Michael Helbling]: Well, they should be nice to me anyways. That’s your, uh, that’s your, like, uh, thing. You have a thing.

00:39:25.18 [Tim Wilson]: Like, we need the Stanford Code. Rock, flag, and Microsoft Excel. Hold on, one more. Rock, flag, and pivot tables.

00:39:37.74 [Michael Helbling]: There you go.

Podcast: Download | Embed

Subscribe: RSS