1. And... we're back. You'll notice that all of your images are back as well, as are our beloved emoticons, including the infamous :foo: We have also worked with our server folks and developers to fix the issues that were slowing down the site.

    There is still work to be done - the images in existing sigs are not yet linked, for example, and we are working on a way to get the images to load faster - which will improve the performance of the site, especially on the pages with a ton of images, and we will continue to work diligently on that and keep you updated.

    Cheers,

    Fok on behalf of the entire Styleforum team
    Dismiss Notice

Australian Members - Part II - if you read the first post, you'll get what this is all about.

Discussion in 'Classic Menswear' started by Foxhound, Feb 10, 2016.

  1. Coxsackie

    Coxsackie Senior member

    Messages:
    3,964
    Joined:
    Sep 30, 2013
    Location:
    Melbourne - Sydney - Shanghai
    @crdb , just wondering - if I shop online at one of these savvy sites using my MacBook, I'm likely to be given a higher price, right? OK. So how about if I use a VPN? Will this hide information regarding my device type?

    I'm thinking, I should start using a VPN with (say) a Filipino server, then reset the currency to A$, or is that giving the game away?

    Anybody with direct experience, preferably successful, using such techniques - please speak up.
     
  2. crdb

    crdb Senior member

    Messages:
    713
    Joined:
    Mar 22, 2014
    I'm going to be brief(er) because this is SF, not HN. That being said, I feel like after a few years and lessons learnt I ought to say this: if you master the computer science as opposed to the toys, in the long run you will have a better career.

    When you say "Spark, Tensorflow" I'm hearing the same hard working young devs say "Hadoop, Map/Reduce" 5 years ago and so on. So first, why are you learning about distributed computing already? How big are your datasets, really? Because by going for distributed you are losing out on an enormous number of things, like provability (with SQL, dealing with data in a relational way, that's the set theory meets predicate logic way, not the ER diagrams way) or the man-centuries of research and libraries and implementations in something like R (or you'll limit yourself to whatever has already been written in the new paradigm).

    In my case I was obsessed with getting neural nets to work on GPUs in Haskell. Never had a use case for the stuff, nor managers willing to put in the budget to basically rewrite entire R libraries from scratch in a parallel, GPU-friendly way. But I was damn sure it would make or break my career (it didn't).

    On the other hand, I meet very few devs - maybe a few more in Australia than the rest of APAC - that are capable of thinking about databases declaratively, which is a great business opportunity for us as we're hired to clean up but which I find slightly sad since it's not hard to figure out. And I keep interviewing people who have "years of [insert big data framework du jour] experience" but can't tell me the difference between outliers and high leverage observations.

    And the folks who can build an [insert framework] process, they build something which looks like work has been done and the smarter ones cherry pick good historical data to show the non-technical management progress and they hope to job hop or move into management before anybody notices. And people rarely notice, because most legacy codebases are giant spaghetti balls that nobody dares to touch in case existing stuff breaks.

    On the same subject: https://scottlocklin.wordpress.com/2015/08/28/advice-to-a-young-social-scientist/

    \rant
     
    Last edited: Jul 6, 2016
  3. crdb

    crdb Senior member

    Messages:
    713
    Joined:
    Mar 22, 2014
    It really depends on sites. I know some big AU sites that aren't bothering, either out of laziness or because they don't have the capability. On the other hand, American sites are more aggressive.

    The best way to know is to try a few devices and figure out what they are detecting. It's usually pretty rough. I've seen prices jump up on the very listing I'm looking at on AirBNB after I log in... that's a major UX fail in my book. For what I buy, switching to a cheap old laptop with an older OS usually does the trick.

    FWIW the Philippines is not necessarily the best idea, since that's the kind of country where a not insignificant number of customers are ultra-rich and the rest are on mobile devices.

    One more reliable way to reduce costs is to sign up to the newsletter of new outfits, since that's where a lot of the marketing money will go (it looks like you're "reactivating" customers and the voucher % is usually lower than the Adwords cost per order, and in some cases isn't even accounted for, so investors love it).
     
    1 person likes this.
  4. Epicure

    Epicure Senior member

    Messages:
    126
    Joined:
    Apr 11, 2016
    Location:
    Australia
    I'm unfortunately very put off trying Meermin based on what I've read about their customer service.

    And the risk of getting sizing wrong would appear quite high. Return shipping costs are steep and potential to resell at a reasonable recovery also seems risky, living in Australia as opposed to the US, for example.

    Regarding the points made about Church's shoes, I would add that it's probably a great time to buy them from Herring if there's something there you like. Haven't worn them before myself, but I wouldn't be opposed to trying them either at current prices.
     
  5. crdb

    crdb Senior member

    Messages:
    713
    Joined:
    Mar 22, 2014
    Ah, I basically assume there'll be no customer service and the shoes are non-returnable (or in other words: at least part of the premium with other brands pays for other people's customer service, which I'm not necessarily keen on). The way I priced it was P(wrong size) * cost which is 20% * $150 = $30, still less than the savings relative to other brands. My time (for the time taken to return the pair by mail) is worth more than the full price of the shoes anyway. 20% is conservative if you measure your foot well and read the thread on sizing - I got it right first time.

    But are there any Australia businesses that sell shoes in even the same order of magnitude price with a Goodyear welted sole?

    Also are Meermins less likely to sell in Australia? I'd have thought there'd be a nice market for the things considering the lack of cheap options locally.
     
    1 person likes this.
  6. The Ernesto

    The Ernesto Senior member

    Messages:
    3,471
    Joined:
    Nov 22, 2010
  7. Pink Socks

    Pink Socks Senior member

    Messages:
    1,708
    Joined:
    Apr 21, 2009
    

    Me too...and I love my pairs too.
     
    1 person likes this.
  8. clayb

    clayb Member

    Messages:
    19
    Joined:
    Jun 27, 2016
    Sorry for coming across like that. I didn't mean to try to learn Big Data stuff via those 'toys', they are just to give me the first taste of this emerging technology so that I can decide if it's for me or not. And I wholeheartedly with you that we should start with the basic and master it first. Personally I have problems with people who like to tell me web development equals to knowing some fancy frameworks like React etc.

    Quote: For some it can be billions of records. I'm not sure if that's big enough or not but the Big Data stack, while is still evolving, has been pretty much decided already.

    Quote: Coincidentally, I have been eyeing the GTX 1080 for some experiments/personal projects, but will have to think through it first.

    Quote: That's a surprise to me. I always thought removing outliers must be one the first things to do. I've been warned that a majority of time would be spent on cleaning up the data, not the interesting stuff like analysing, training, or modelling.

    Quote:
    I have experienced it :) C++ classes with thousands of lines and people keep adding new methods to them and no one dares, or bothers, to refactor the 20 year old code base. It's a text (e.g. NoSQL) database btw.

    Quote: Interesting link. Also on the same subject, I know someone, who has a background in statistics, always complains to me that they don't have enough of knowledge of computer science to do the job good enough and I always have to assure them that their own knowledge is more important and that they only need to know the basic of CS to run commands and use the tools provided etc.

    For me I guess I'm still at the exploration phase, pondering which area to dive into. Australia, imho, is not a very large market when it comes to IT, not to mention the outsourcing trend, and it's important to choose the right areas to stay relevant and competitive in the long term. Big Data seems to be one of them, and also is intellectually challenging, which is attractive to me. So I decide to give it a shot and will see how it goes.
     
    Last edited: Jul 7, 2016
  9. clayb

    clayb Member

    Messages:
    19
    Joined:
    Jun 27, 2016
    Thanks Coxsackie, it sounds like a lot of work to be done when you get caught in a heavy rain with your leather shoes. I'm going to topy most of my shoes, except maybe the C&J pair, to minimise the risk, but will probably also get a pair of waterproof boot or something similar with the sole types suggested.

    There is Belmore. They happen to be having a sale right now too. I've never bought anything from them so am not sure if their shoes can be compared with Meermin or not, but they claim that the leather is full grain kangaroo.

    Quote: I'm afraid if they come to Australia, their price will not be as good as of now. We can just look at Loake.
     
  10. Oli2012

    Oli2012 Senior member

    Messages:
    2,267
    Joined:
    Aug 22, 2012
    Location:
    Sydney, Australia
    Anyone keen on two of Jason's knits?

    [​IMG]

    Unworn. $75 for both.
     
  11. Coxsackie

    Coxsackie Senior member

    Messages:
    3,964
    Joined:
    Sep 30, 2013
    Location:
    Melbourne - Sydney - Shanghai
    Not really. I was out and about last night in Melbourne, with intermittent drizzle and plenty of water on the ground, in leather-soled shoes. I simply avoided stepping in puddles. Shoes were fine.

    I won't get into the Topy argument (which has been played out to the death on this thread many times over), except to say that they are not my personal preference.

    Bottom line: when building your shoe collection, try to end up with at least one pair of black and one of brown fitted with some kind of waterproof sole. Check the weather report before deciding on your footwear for the day. But don't panic if you find yourself caught out in the rain with leather soles - just take a bit of extra care.
     
    3 people like this.
  12. eightace

    eightace Senior member

    Messages:
    333
    Joined:
    Jul 29, 2011
  13. md2010

    md2010 Senior member

    Messages:
    1,894
    Joined:
    Jul 11, 2011
    Location:
    Sydney
    Great deal actually.
    What is everyone's thoughts on a Navy M65 jacket ? or it must be olive ??
    Application - casually on the weekend & over suit not so often.
     
    Last edited: Jul 7, 2016
  14. crdb

    crdb Senior member

    Messages:
    713
    Joined:
    Mar 22, 2014
    Billions of rows is standard for an analytical workload. You're going to have to tweak your DB a little bit to speed up ELT and be careful with indices and maybe a bit of partitioning, but you can stay relational and you might even be able to do it on your laptop if it's got enough disk. Google "Postgres billions of rows" and you'll find some threads dating back to the early 2000s when hardware was a lot less capable... If you haven't yet, I'd recommend reading an easy book like https://www.amazon.com/SQL-Antipatterns-Programming-Pragmatic-Programmers/dp/1934356557 as well as a proper theoretical one like https://www.amazon.com/Introduction-Database-Systems-8th/dp/0321197844/r or https://www.amazon.com/Relational-Model-Database-Management-Version/dp/0201141922/.

    I'd shy away from GPUs unless you are interested in low level programming. Truth is, there's very few genuine applications for GPUs - maybe very complex algorithms running on very large datasets, or stuff like... video games actually processing graphics - so it's a bit of going down the rabbit hole with not much to show for it. If you're interested in parallel programming this is a decent book: http://chimera.labs.oreilly.com/books/1230000000929/index.html

    Cleaning the data IS the interesting part. Because that's one of the things that has the largest impact on the performance of your model. It forces you to make a lot of design decisions. Model selection, validation, etc. are fun the first few times, but I'd argue a well designed multilinear regression is going to beat deep nets and XGBoost if the guys building it know what they are doing. In fact they have in a recent Kaggle competition which created quite a bit of buzz in the community.

    You don't just "remove outliers". Some observations naturally have a lot of influence on the model but they may be valid. In which case you need to do some transformations on your variables or you need to pick a different model. E.g.
     
  15. sliq

    sliq Senior member

    Messages:
    1,086
    Joined:
    Apr 19, 2012
    Location:
    Sydney, Australia
    get both :D
     
    2 people like this.
  16. Pink Socks

    Pink Socks Senior member

    Messages:
    1,708
    Joined:
    Apr 21, 2009
    @crdb - all very interesting stuff from what I can follow. Any recommendations for a real beginner to look at just to get an overview and understanding - website, book, video. Just enough so I can keep up with this thread? (Seriously, although I feel like I should add a menswear comment too - what is appropriate attire for analysing big data and computer programming - Cucinelli cashmere hoodie?)
     
    1 person likes this.
  17. crdb

    crdb Senior member

    Messages:
    713
    Joined:
    Mar 22, 2014
    I think the two most important things are:
    - understanding the relational model, and applying it in the best open source database available today, PostgreSQL (the most relational of the lot by far thanks to decades of academic research in Ingres and elsewhere);
    - understanding statistical learning and if you have time, statistics itself.

    The first allows you to reason about data declaratively - that is, without specifying how whatever you like is computed. It's actually incredibly conceptually easy; if you understand Venn diagrams and logic, you can write correct, relational SQL. Which is why I am mystified that most CS courses teach it from a flawed POV that dates from before Codd's seminal paper (https://www.seas.upenn.edu/~zives/03f/cis550/codd.pdf).

    So, you take your input data, and you literally "declare" what you want and voila, data cleaned and result obtained, provably correctly. Understand the concept of a transaction, of a logical unit of work, of a relation variable vs a relation (a variable vs values, if you will), domains, types, etc. and you're good to go.

    The second is about making sense of the data. Curve fitting, basically. Your brain does it everyday with everything, processing GB of data per second from all your senses. Reading SF, you're looking at posts about clothes from dressers of varying ability; first you "learn" which ones are good, then you "learn" why what they are doing is good, and voila, you have learnt about fashion by abstracting from examples (which are your data). There might be obvious patterns (the X vs )( quarters, the jacket ending halfway your silhouette, the famed "Northern Lights"), constraints (no open lacing with a suit) and non-obvious ones ("which skin type works best with which shirt pattern and colour palette").

    Statistical learning is the formalisation of this. You have data, you fit a model to it (by minimising the error between the model and the data, usually) and you derive some kind of use from it (in the SF example: you learn to dress "better" although really you learn to buy expensive clothes that very few people will understand beyond "he looks nice"). You can use these models for intuition (e.g. aforementioned "obvious patterns" that "explain"; or the revenue equation mentioned before) or for prediction (try a bunch of new shirt and jacket patterns together and "feel" that they are wrong or right, i.e. the amount of "error" in what you just did vs what you think looks good, which you could call taste).

    And there we talk about the separation between model and implementation. The model is only concerned with how things are, defining your input and output, at a conceptual level. Implementation is about how you make it happen. This is a very important distinction. A constraint on an SQL column is a model consideration: this column can only take these values, how you implement it is not important but it has to happen. An index is an implementation consideration (although something like CREATE UNIQUE INDEX WHERE [logical statement]; in SQL straddles the two - it's an implementation trick used to implement a model-level constraint). From that point of view, statistical learning is about the model, and Spark/Hadoop/Redshift (yes you can)/R/whatever is about implementation (at different levels).

    I used to try and learn from MOOC but in my experience you just pick up patterns of behaviour that you can then apply in a job without really understanding the fundamentals. That used to cut it in 2008, not so much today. For the same number of hours, read the textbooks and understand what they say, then be able to abstract from that to new situations, patterns and models, and you're a much better thinker for it. A CM equivalent might be the difference between understanding the reason for which a wool tie does not go with the finest worsted suit, or understanding why things work at different levels of formality, vs parroting "no brown in town".

    And so I repeat my recommendations: ISLR for statistical learning (free on http://www-bcf.usc.edu/~gareth/ISL/ - although you can bump up to ESLR if you feel comfortable with linear algebra) and Code or Date for the relational model as per above post. Date is I think a bit more readable. They disagree on a few issues. Total reading time 20-50 hours depending on how comfortable you want to get with the material.
     
  18. crdb

    crdb Senior member

    Messages:
    713
    Joined:
    Mar 22, 2014
    I should probably point out that I'm the business guy of the company so take what I say with a pinch of salt :p
     
  19. The Ernesto

    The Ernesto Senior member

    Messages:
    3,471
    Joined:
    Nov 22, 2010
    

    I would have thought something with a pattern.
     
  20. crdb

    crdb Senior member

    Messages:
    713
    Joined:
    Mar 22, 2014
    [​IMG]
     
    3 people like this.

Share This Page

Styleforum is proudly sponsored by