分析工作試用期收獲

Have you been hearing the new industry buzzword — Data Analytics(it was AI-ML earlier) a lot lately? Does it sound complicated and yet simple enough? Understand the logic behind models but don't know how to code? Apprehensive of spending too much time learning to code before jumping on the bandwagon?

您最近是否經(jīng)常聽到新的行業(yè)流行語-Data Analytics( 早于AI-ML) ？ 聽起來復(fù)雜但足夠簡單嗎？ 了解模型背后的邏輯，但不知道如何編碼？ 擔(dān)心在投入潮流之前花太多時間學(xué)習(xí)編碼嗎？

Worry not, there are some awesome tools available for free for non-coders that can help develop complicated models in no time. These tools are completely free for personal use, extremely easy and intuitive and can help one practice without the hassle of learning how to code.

不用擔(dān)心，有一些很棒的工具可供非編碼器免費使用，這些工??具可以立即幫助開發(fā)復(fù)雜的模型。 這些工具完全免費供個人使用，非常簡單直觀，可以幫助一種實踐，而無需學(xué)習(xí)如何編寫代碼。

I am an amateurish coder but a big machine learning enthusiast. I can code but I avoid it as much as I can (Thank God for that Recording Macro option in Excel), till the point I cannot avoid it.

我是一個業(yè)余編碼員，但是非常喜歡機器學(xué)習(xí)。 我可以編寫代碼，但我會盡量避免(感謝上帝，感謝Excel中的那個Recording Macro選項)，直到無法避免為止。

I was working on developing a model for forecasting traffic on a road and had to try a lot of things when I started looking for non-coder resources and found these gems. I am discussing the best three I found. Again, these are open source for individual users but have priced versions for commercial uses.

我當(dāng)時正在開發(fā)一種用于預(yù)測道路交通量的模型，當(dāng)我開始尋找非編碼器資源并發(fā)現(xiàn)這些寶石時，不得不嘗試很多事情。 我正在討論我發(fā)現(xiàn)的最好的三個。 同樣，這些是面向個人用戶的開源軟件，但是具有商業(yè)用途的定價版本。

這些工具不能做什么 (What These Tools Cannot Do)

Please be aware, although these tools remove the need for coding, your understanding of models, basics of data preparation, and statistics should be above the bare minimum. The reason is that when you code, you exactly know what is being done and how, while in most of these tools, default parameters are preloaded, and sometimes the code is not visible to the user. Thus it is easy for model errors to go unnoticed in case the user does not do a thorough QA.

請注意，盡管這些工具消除了對編碼的需求，但是您對模型的理解，數(shù)據(jù)準(zhǔn)備的基礎(chǔ)知識和統(tǒng)計信息應(yīng)該高于最低要求。 原因是在編寫代碼時，您確切地知道正在執(zhí)行的操作以及如何執(zhí)行操作，而在大多數(shù)這些工具中，默認(rèn)參數(shù)是預(yù)加載的，有時代碼對用戶不可見。 因此，如果用戶沒有進(jìn)行全面的質(zhì)量檢查，很容易引起模型錯誤的注意。

In addition to this, these tools will not tell you which data cleaning technique to use, which model to build, or which statistic to compare instead, the tools will let you do all the above tasks easily and give you more time to think and analyze data.

除此之外，這些工具不會告訴您使用哪種數(shù)據(jù)清除技術(shù)，要構(gòu)建哪種模型或要比較哪種統(tǒng)計量，這些工具將使您輕松地完成上述所有任務(wù)，并給您更多的時間進(jìn)行思考和分析數(shù)據(jù)。

Now that you have read all the warnings let us directly dive in.

現(xiàn)在您已經(jīng)閱讀了所有警告，讓我們直接潛入。

1. Knime Analytics (1. Knime Analytics)

This is by far, the best tool in the open source domain.

到目前為止，這是開源領(lǐng)域中最好的工具。

Knime is a very intuitive platform that helps create models using drag and drop nodes in a workflow kind of environment. It is built on python, has widgets for data input, data cleaning, modeling (regression, clustering, classification, Neural Networks, etc), statistics, and majorly used representations.

Knime是一個非常直觀的平臺，可在工作流環(huán)境中使用拖放節(jié)點幫助創(chuàng)建模型。 它基于python構(gòu)建，具有用于數(shù)據(jù)輸入，數(shù)據(jù)清理，建模(回歸，聚類，分類，神經(jīng)網(wǎng)絡(luò)等)，統(tǒng)計信息和主要使用的表示形式的小部件。

It is has a desktop version (I love it) and a Server version for people who want to develop and deploy these model workflows on the web. Installing Knime on your machine is fairly easy, and using it is even more. Below is an example of an NN Model.

它有一個臺式機版本( 我喜歡它 )和一個服務(wù)器版本，供希望在網(wǎng)絡(luò)上開發(fā)和部署這些模型工作流的人們使用。 在您的計算機上安裝Knime非常容易，使用它甚至更多。 以下是NN模型的示例。

There are nodes for every action needed to build a Neural Network. Importing the data, partitioning it, feeding a part to a learner, a predictor (test set), and then a scorer for checking the accuracy of the model. Parameters can be set in nodes that are connected to each other using connectors and can be executed in sequence.

建立神經(jīng)網(wǎng)絡(luò)所需的每個動作都有節(jié)點。 導(dǎo)入數(shù)據(jù)，對其進(jìn)行分區(qū)，將零件饋給學(xué)習(xí)者，預(yù)測變量(測試集)，然后饋給評分員以檢查模型的準(zhǔn)確性。 可以在使用連接器相互連接的節(jié)點中設(shè)置參數(shù)，并且可以依次執(zhí)行。

Credits — Knime Workspace on my Desktop積分-我桌面上的Knime工作區(qū)

2.橙色 (2. Orange)

Orange is an open source machine learning, data visualization, and analysis tool. Orange also works on widgets arranged in a workflow pattern and has some specialized libraries for specific tasks (time series, bioinformatics, etc).

Orange是開源的機器學(xué)習(xí)，數(shù)據(jù)可視化和分析工具。 Orange還可以處理按工作流程模式排列的小部件，并具有一些用于特定任務(wù)(時間序列，生物信息學(xué)等)的專用庫。

Orange’s UI is more fluid but its node list is less exhaustive than Knime. It has numerous visualization options and can produce decent data analytics. It is built on python and can help create and evaluate models for regression, classification, NN, clustering, time series among other things.

Orange的UI更加流暢，但其節(jié)點列表不如Knime詳盡。 它具有多種可視化選項，可以進(jìn)行體面的數(shù)據(jù)分析。 它基于python構(gòu)建，可以幫助創(chuàng)建和評估模型以進(jìn)行回歸，分類，NN，聚類，時間序列等。

Orange Website-Orange網(wǎng)站

3.藍(lán)天統(tǒng)計 (3. BlueSky Statistics)

Bluesky is an R based tool that can be used for data modeling and visualizations. It is open source and available for desktops. It has a rich GUI and it can help ease the learning curve for R newbies as for each function the R code is visible.

Bluesky是基于R的工具，可用于數(shù)據(jù)建模和可視化。 它是開源的，可用于臺式機。 它具有豐富的GUI，它可以幫助R新手簡化學(xué)習(xí)過程，因為R代碼可見的每個功能。

BlueSky lacks workflow style architecture & node functionality. Instead, it has functions listed under tabs similar to MS Office ribbon tabs. The beauty of BlueSky is that it is built on R which is an incredibly powerful language for statistical data analysis. It has command editor and as the code is completely visible to the user, it is extremely easy for users to modify the code as they like it. It ensures that regular users of R can save a considerable amount of time using this application.

BlueSky缺乏工作流樣式的體系結(jié)構(gòu)和節(jié)點功能。 相反，它具有類似于MS Office功能區(qū)選項卡的選項卡下列出的功能。 BlueSky的優(yōu)點在于它基于R，R是一種用于統(tǒng)計數(shù)據(jù)分析的功能強大的語言。 它具有命令編輯器，并且由于代碼對用戶完全可見，因此用戶可以輕松地隨意修改代碼。 它確保R的普通用戶可以使用此應(yīng)用程序節(jié)省大量時間。

Credits — BlueSky Stats User Manual積分-BlueSky Stats用戶手冊

There are numerous data analytics tools available in the market but most of them are not open source. This makes it difficult for individual users who are still in the exploratory phases of data science.

市場上有許多數(shù)據(jù)分析工具，但是其中大多數(shù)不是開源的。 這使得仍處于數(shù)據(jù)科學(xué)探索階段的個人用戶很難。

These three tools are my top favorite to dabble with small Data Analytics problems. They can save an immense amount of time for newbies who might be daunted with the idea of learning to code.

這三個工具是我最喜歡的小數(shù)據(jù)分析問題。 對于那些可能對學(xué)習(xí)編碼的想法望而卻步的新手來說，它們可以節(jié)省大量時間。

This list is based on tools available in late 2019. I will update this if I find any more similar tools. I hope you find this story helpful in beginning your journey into Data Analytics!

該列表基于2019年末可用的工具。如果我發(fā)現(xiàn)更多類似的工具，我將對其進(jìn)行更新。 我希望您發(fā)現(xiàn)這個故事對您開始數(shù)據(jù)分析之旅有所幫助！

翻譯自: https://towardsdatascience.com/explore-data-analytics-with-zero-coding-skills-for-free-f2c982d1e2d6

分析工作試用期收獲

總結(jié)

以上是生活随笔為你收集整理的分析工作试用期收获_免费使用零编码技能探索数据分析的全部內(nèi)容，希望文章能夠幫你解決所遇到的問題。

如果覺得生活随笔網(wǎng)站內(nèi)容還不錯，歡迎將生活随笔推薦給好友。