Тёмный

Alteryx vs. KNIME (vs. Excel) in 2021: Which to use for analysis in 

Eric
Подписаться 1,7 тыс.
Просмотров 6 тыс.
50% 1

There are several new, and very powerful, analytics programs that can take the place of Excel. Should they, and which one should you use?
Watch this video to understand the differences between KNIME and Alteryx as well as why you should consider using them instead of Excel in 2021.
My Udemy course on KNIME: www.udemy.com/course/data-sci...

Хобби

Опубликовано:

 

31 июл 2024

Поделиться:

Ссылка:

Скачать:

Готовим ссылку...

Добавить в:

Мой плейлист
Посмотреть позже
Комментарии : 31   
@thiemokellner1893
@thiemokellner1893 7 месяцев назад
Thanks! Doing KNIME courses right now (they are from KNIME, are good, low cost or even free). You can also get certified in KNIME. So my comments will be for KNIME only. I might have understood it wrong, but while you can tell KNIME to execute any arbitrary node, as long as it and its predecessors are configured. It means, however, that any preceding node not having been executed yet, gets executed. The possibility to execute stepwise is very, very compelling to me. With respect to data exploration, with the new interface of KNIME you do not even need to open a window by context/shortcut. Select the node of interest and the data, as long as the node has been executed, and a table at the bottom shows the data as it looks after the processing of the respective node. So far, I have felt the KNIME community helpful and good. There are also KNIME developers in them. I suspect, they get a portion of their worktime to help people there. About the costs: KNIME is a for-profit organisation, too. However, it is true that their Analytics platform is free and as far as I understood open source. If you only want to use your workflows/report on your local computer, you should be fine with it. However, if you want to automate the generation of reports (e.g. pdf), charts (e.g. web based) etc. you probably need their non-free business hub.
@davidlopez-fe2lb
@davidlopez-fe2lb 3 года назад
Hi Eric, great overview of the two tools...really enjoyed it. As an avid Alteryx user I wanted to add some perspective and maybe get your feedback on them. Data Exploration: You mentioned you can't disable part of your workflow you have to run the entire flow, this can be fixed by putting your flow into tool containers and disabling them allowing you to run pieces or all of your workflow. I came across this issue when I first started us Alteryx and was frustrated until I found out about the 'tool container tool'. You also said you have to add a browse tool after every process to see the output, this isn't entirely true. You can view the output without ever using a browse tool if you just click on the green arrow after the tool, this will show you what the output looks like at every step; but you're correct in that you can see details about the data output, you'll only see it in a data table format. Data Cleaning: you mentioned KNIME 'guesses' the data types a lot better than Altyerx, that's because Alteryx doesn't try to 'guess' your data, it reads it in it's native form. If you did want Alteryx to 'guess' your data, just drop an Auto Field tool in and it does what I imagine KNIME does. I wouldn't see it as a positive or negative, one guesses and one doesn't, depending on your need that could be great or not so great feature. Flexibility: KNIME has 'add ins', Alteryx has the same concept in the Alteryx Gallery. Alteryx is out of the box, but you can customize that box by downloading 'tools' from the Alteryx gallery. Think of it as importing packages in R, same concept with Alteryx and it's free! I will say one weak point of Alteryx is APIs, as far as connections go, you have to be a borderline Full Stack developer to get APIs working in Alteryx, which isn't a great model for a 'no code' tool. ML Models: I use ML models in Alteryx quite frequently, and haven't had any problems but I don't work with 'big data' as much as my counterparts; so I couldn't accurately speak to the model size limitations. I'd also suggest the "Machine Learning" ribbon, it provides an even more clear cut approach to ML models for 'citizen data scientists' but it does incur an extra fee for access to that ribbon. Community: I haven't used KNIME community, but I'd be absolutely AMAZED if their community is better than Alteryx is as you mentioned. Their community is on par with Tableau community which I think should be the standard for tool communities. Price: Of course I can't argue for Alteryx on this, I'm lucky enough to work with a company that provides access to Alteryx so price isn't an issue. I'm happy I came across this video, and I'm more than happy to check out KNIME because clearly I'm in the Alteryx band but am open to learning about KNIME. Thanks again. Cheers
@ewhulbert
@ewhulbert 3 года назад
Whew, lot going on here. Tool containers: I thought I mentioned tool containers? Anyways, you still have to run the whole container, not nearly as nice as KNIME when you can run node by node. Browse: holy crap, has it always been like that? Why the hell do they have the browse node then? It seems completely redundant. Auto-field: I did not know that, I'll need to check that out. Gallery: From what I saw, it was not as useful as the myriad of add-ins for KNIME. Perhaps I missed them, and I think my org might actually have gallery disabled for infosec purposes. ML: I do have the ML ribbons and it simply is not as good. Try K-Means between the two, there is a difference. Community: I thought the alteryx community was a tad artificial with what looked like sponsored posts or people who were asked to load stuff. Hard to beat the community of a free tool vs. paid. I don't by any means hate Alteryx. I use it frequently.
@davidlopez-fe2lb
@davidlopez-fe2lb 3 года назад
@@ewhulbert Yah that's true, so if you have four tools in your container - but maybe 17 tools in your entire workflow - and you want to run just one of those four you'd have to run that whole container, although you could drop containers in containers but that's a bit of a mess. I do enjoy R where you can highlight a specific part of your script and run just that piece, wish Alteryx would introduce a feature like this. Yup, it's been like that for a while, pretty neat feature. I do want to emphasize that it doesn't have the features the browse tool does, again you just see a table output when you click on that little green arrow, but with the browse tool you can see a table view but also do a lot more. I typically check for data quality initial with a browse tool. Yah it's a great feature, Alteryx doesn't always read '01/14/2009 12:25' as a datetime type, especially if your data is dirty, but run it through Auto field and it'll fix that. There's quite a bit of free tools out there, and like yourself, my org does but some restrictions on what I can and can't download. I'll have to check that out, I do clustering quite a bit as wheel. I got it, yah it does look a little like a super bowl ad at times, but it's super helpful. Almost belongs in the "Support" section, when I get stuck with Alteryx I go to the community first and then my account manager second. I rarely have to go to my account manager because the support the community provides it great. Thanks again for the content, I'll have to learn more about KNIME which is exciting. cheers
@ewhulbert
@ewhulbert 3 года назад
@@davidlopez-fe2lb I can't hate on Alteryx for not reading dates / datetimes right the first time. In my experience, no platform is particularly good at it. KNIME certainly isn't for dates only, does really good for datetime but non standard date formats trip it up and I have to manually input date formats.
@the_learner12
@the_learner12 3 года назад
If you learn how to use KNIME, it is excellent. What blows my mind is that it is free.
@ewhulbert
@ewhulbert 3 года назад
Yeah, KNIME is really impressive and its incredible that its free.
@FelipeDSFranca
@FelipeDSFranca 3 года назад
Being an Alteryx user for quite some time, I just found out about Knime and I'm in love with it.
@bobreddert7547
@bobreddert7547 2 года назад
I'm in the middle of a project migrating Alteryx workflows to Knime as a cost-saving measure (Alteryx is quite pricy). This review is excellent and I completely agree with most of your points. However, I think you missed a category for evaluation: performance. When reading from local files (spreadsheets, delimited, etc.) Alteryx is definitely faster than Knime and if you use Alteryx's proprietary Alteryx DB format, it blows Knime away.. However when pulling data from relational databases, we tested 5 different database platforms and in every case, Knime was much quicker.
@ewhulbert
@ewhulbert 2 года назад
I hadn't noticed a speed difference from spreadsheets and I have used some very large ones. Perhaps I should go back and look. Good point on the Alteryx DB format. It really shrinks sizes down to and is quite fast. To me, for enterprise applications, I'd much rather have speed from relational databases than for local files, so overall I'd chalk it up as a win for KNIME. Thanks for your insight, and best of luck with your transition.
@henrique4171
@henrique4171 2 года назад
Do u think that i should use python for high storage enviroments,or i can use in these?
@causa04
@causa04 3 года назад
power query can do a lot of this things that you mentioned excel can't do or it is slow for.
@ewhulbert
@ewhulbert 3 года назад
It can, but it still struggles with size and doesn't have the visual workflow to keep track of complicated multi-step processes.
@diegocardenaguillen4312
@diegocardenaguillen4312 Год назад
Great information but at times it was hard to understand what uou
@Palantir_Daily
@Palantir_Daily 3 года назад
In your opinion, how does Palantir’s no code tools compare?
@ewhulbert
@ewhulbert 3 года назад
I really don't know. The hard part with things like that is they are selling a whole package, so there is not really a way to try it out without joining an organization that uses it. They are pretty vague in the descriptions on their site as to how it works, and I don't see any way that a hobbyist can get smart on it. One of the reasons Python and R took off, in addition to them being easy languages, was anyone could use them, learn to use them, and build things for them. KNIME is similar to that. Ultimately what matters if you're an organization is "can I hire people who know how to use this?" That is why AWS, Azure, etc., are all enabling all sorts of functionality from other languages so with minimal training any competent analyst or data scientist can be up and running quickly. Its not apparent to me if Palantir is the same way or not.
@Palantir_Daily
@Palantir_Daily 3 года назад
@@ewhulbert Thank you for taking the time to share! I ask because of the recent IBM & Palantir contract. I believe IBM said they are entering the agreement to gain access to Palantir’s no code AI. Must be decent for IBM to partner and simply use Palantir’s no code AI rather than develop their own. So I was curious if you had an opinion! Thanks for giving insight.
@AlexanderBibin
@AlexanderBibin 3 года назад
Nice overview, however I disagree with your assessment that Alteryx can't be extended beyond out-of-the-box functionality. People build macros and apps with alteryx tools to extend it wide and far, while more technically apt users use Alteryx' SDKs to build custom tools in Python (and soon in other languages as well).
@ewhulbert
@ewhulbert 3 года назад
Sorry if I gave the impression that it cannot be extended. It most certainly can; my company has some proprietary ones that we have created to help our teams. However, it is far less frequent than an open source tool like KNIME where the goal is sharing information and capabilities.
@AlexanderBibin
@AlexanderBibin 3 года назад
@@ewhulbert Perhaps it is less frequent than in Knime (not that familiar with it), but Alteryx community is very thriving and responsive. There are also lots of publicly available extensions built both by Alteryx folk themselves and by the community in their gallery.
@kaegancasey5291
@kaegancasey5291 2 года назад
Managing macros sucks though compared to KNIME components it's way easier to extend knime and create reusable building blocks of logic than Alteryx. Also you can't have more than one macro with a python tool in it since they both use the same temporary directory which is just Alteryx shooting itself in the foot for creating any sort of more advanced re-usable functionality
@rose8968
@rose8968 3 года назад
Alteryx community is amazing
@ewhulbert
@ewhulbert 3 года назад
Glad you like it.
@HaiLeQuang
@HaiLeQuang 3 года назад
I wonder what is your day job that helped you gain that much exposure with analytics tools?
@ewhulbert
@ewhulbert 3 года назад
I'm a management consultant. Use Alteryx in my day job as my firm has licenses through them. Have used KNIME in past jobs, and occasionally in this one for turning over work to a client who does not have an Alteryx license.
@rose8968
@rose8968 3 года назад
Alteryx has tool container that you don’t have to run the entire workflow each time
@ewhulbert
@ewhulbert 3 года назад
To be honest, the tool container sucks; not in the same league as KNIME being able to run node by node. The reason I say this is two fold: 1. It doesn't keep active data in there. If you have a multi-step process you need to create output files and have them at the end of your container because the moment you disable the tool container to not run it gets rid of the active data. This is massively frustrating but a necessity. I do it all the time because with huge data sets it would otherwise take forever to make small changes. 2. You have to fully disable it to get it to not run, and when you do that you cannot see what was in it and what the outputs were at various points. The tool container is more akin to KNIME's metanode and there really is not a good equivalent to being able to run node by node.
@FedericoLov
@FedericoLov 2 года назад
I'm sick and tired of people talking about excel while ignoring powerquery and powerpivot
Далее
Документы для озокомления😂
00:24
вернуть Врискаса 📗 | WICSUR #shorts
00:54
Это конец... Ютуб закрывают?
01:09
3 Essential Excel skills for the data analyst
18:02
Просмотров 1,5 млн
Airflow Vs. Dagster: The Full Breakdown!
14:51
Просмотров 6 тыс.
Three Best AI tools for Data Analysis
15:39
Просмотров 30 тыс.
Learn Data Science Using KNIME by Scott Fincher
1:06:13
KNIME tutorial: Grouping data with the GroupBy node
7:48
Best exercises to lose weight ! 😱
0:19
Просмотров 9 млн