<![CDATA[Nur Fadhilah]]>https://nurfadhilah.techRSS for NodeTue, 07 Apr 2026 20:44:03 GMT60<![CDATA[8 tips when building a high fidelity prototype in Figma]]>https://nurfadhilah.tech/8-tips-when-building-a-high-fidelity-prototype-in-figmahttps://nurfadhilah.tech/8-tips-when-building-a-high-fidelity-prototype-in-figmaTue, 10 May 2022 21:15:52 GMTResources:

Figma's documentation of best practices

Figma's Community Page

  1. Do browse the community page for UI kits to quickly bootstrap your project.

  2. Decide early in the project the color palette and fonts to facilitate efficient collaborative design work.

  3. Understand what is a group and what is a frame. Frames and groups are containers in figma but behave differently when resizing.

  4. Shift + A. It easily autoformats the contents for your figma prototype and when you realise you need to overlay other designs, simply delete autoformat. Autoformat has both horizontal and vertical scrolls.

  5. When building, make sure you untick clip content for that frame so that you can comfortably design everything.

  6. A layout grid is useful for alignment.

  7. Build components, so that you can easily duplicate and autoformat to create multiple copies such as in the case of listviews.

  8. Prototype is useful for presentations so break down the user flows according to use cases for early stages. It is easier to edit and then compile all the iterations into final product.

]]><![CDATA[Section 15: Principal Component Analysis (PCA)]]>https://nurfadhilah.tech/section-15-principal-component-analysis-pcahttps://nurfadhilah.tech/section-15-principal-component-analysis-pcaSat, 02 Oct 2021 18:51:10 GMTReferences:

21_ml_15_3Oct21.png

]]><![CDATA[Section 14: Density-Based Spatial Clustering of Applications with Noise (DBSCAN)]]>https://nurfadhilah.tech/section-14-density-based-spatial-clustering-of-applications-with-noise-dbscanhttps://nurfadhilah.tech/section-14-density-based-spatial-clustering-of-applications-with-noise-dbscanSat, 02 Oct 2021 01:21:38 GMTReferences:

20_ml_14_2Oct21.png

]]><![CDATA[Section 13: Hierarchical Clustering]]>https://nurfadhilah.tech/section-13-hierarchical-clusteringhttps://nurfadhilah.tech/section-13-hierarchical-clusteringFri, 01 Oct 2021 01:17:45 GMTReferences:

19_ml_13_1Oct21.png

]]><![CDATA[Section 12: K-Means Clustering]]>https://nurfadhilah.tech/section-12-k-means-clusteringhttps://nurfadhilah.tech/section-12-k-means-clusteringThu, 30 Sep 2021 01:09:35 GMTSo far we have done supervised learning. The remaining sections will be on unsupervised learning. Below is a quick guide on how to pick the estimator:

scikit-learn-algo.jfif Source: scikit-learn

Unsupervised Learning: (1) Clustering: Using features, group together data rows into distinct clusters (2) Dimensionality Reduction: Using features, discover how to combine and reduce into fewer components

Supervised learning's performance metrics (RMSE or Accuracy) will not apply for unsupervised learning.

References:

18_ml_12_30Sep21.png

]]><![CDATA[Section 11: Naive Bayes & Natural Language Processing (NLP)]]>https://nurfadhilah.tech/section-11-naive-bayes-and-natural-language-processing-nlphttps://nurfadhilah.tech/section-11-naive-bayes-and-natural-language-processing-nlpWed, 29 Sep 2021 09:36:04 GMTReferences:

17_ml_10_30Sep21.png

]]><![CDATA[Section 10: Random Forests & Boosted Trees]]>https://nurfadhilah.tech/section-10-random-forests-and-boosted-treeshttps://nurfadhilah.tech/section-10-random-forests-and-boosted-treesMon, 27 Sep 2021 16:03:27 GMTSections 9 and 10 are on tree-based methods. There are three main methods:

Each of these methods stems from the basic decision tree algorithm. Fundamentally, tree-based methods rely on the ability to split data based on information from features. Require a mathematical definition of information and the ability to measure it.

Classification and Regression Tree (CART) introduces many concepts:

  • Cross validation of Trees
  • Pruning Trees
  • Surrogate Splits
  • Variable Importance Scores
  • Search for Linear Splits

References:

16_ml_10_2829Sep21.png

]]><![CDATA[Section 9: Decision Trees]]>https://nurfadhilah.tech/section-9-decision-treeshttps://nurfadhilah.tech/section-9-decision-treesMon, 27 Sep 2021 15:57:57 GMTSections 9 and 10 are on tree-based methods. There are three main methods:

Each of these methods stems from the basic decision tree algorithm. Fundamentally, tree-based methods rely on the ability to split data based on information from features. Require a mathematical definition of information and the ability to measure it.

Classification and Regression Tree (CART) introduces many concepts:

  • Cross validation of Trees
  • Pruning Trees
  • Surrogate Splits
  • Variable Importance Scores
  • Search for Linear Splits

Limitations of a single decision tree:

  • Single feature for root node
  • Splitting criteria can lead to some features not being used
  • Potential for overfitting to data

References:

15_ml_9_27Sep21.png

]]><![CDATA[Section 8: Support Vector Machines]]>https://nurfadhilah.tech/section-8-support-vector-machineshttps://nurfadhilah.tech/section-8-support-vector-machinesSat, 25 Sep 2021 22:18:29 GMTReferences:

13_ml_8a_26Sep21.png 14_ml_8b_26Sep21.png

]]><![CDATA[Section 7: K Nearest Neighbors (KNN)]]>https://nurfadhilah.tech/section-7-k-nearest-neighbors-knnhttps://nurfadhilah.tech/section-7-k-nearest-neighbors-knnSat, 25 Sep 2021 15:46:05 GMTReferences:

12_ml_7_25Sep21.png

]]><![CDATA[Section 6: Logistic Regression]]>https://nurfadhilah.tech/section-6-logistic-regressionhttps://nurfadhilah.tech/section-6-logistic-regressionThu, 23 Sep 2021 19:41:54 GMTReferences:

11_ml_6_24Sep21.png

]]><![CDATA[Section 5: Feature Engineering & Data Preparation]]>https://nurfadhilah.tech/section-5-feature-engineering-and-data-preparationhttps://nurfadhilah.tech/section-5-feature-engineering-and-data-preparationThu, 23 Sep 2021 15:58:34 GMTReferences:

10_ml_5_23Sep21.png

]]><![CDATA[Section 4: Linear Regression (3-4 of 4)]]>https://nurfadhilah.tech/section-4-linear-regression-3-4-of-4https://nurfadhilah.tech/section-4-linear-regression-3-4-of-4Wed, 22 Sep 2021 15:04:47 GMTReferences:

8_ml_4c_22Sep21.png

9_ml_4d_22Sep21.png

]]><![CDATA[Section 4: Linear Regression (1-2 of 4)]]>https://nurfadhilah.tech/section-4-linear-regression-1-2-of-4https://nurfadhilah.tech/section-4-linear-regression-1-2-of-4Mon, 20 Sep 2021 16:00:05 GMTReferences:

6_ml_4a_21Sep21.png

7_ml_4b_21Sep21.png

]]><![CDATA[Capstone Project: Fandango ratings in 2015]]>https://nurfadhilah.tech/capstone-project-fandango-ratings-in-2015https://nurfadhilah.tech/capstone-project-fandango-ratings-in-2015Mon, 20 Sep 2021 15:59:03 GMTToday, I consolidated my understanding by doing a capstone project analysing Fandango's movie ratings in 2015 compared to three other movie aggregators sites—Metacritic, IMDB, and Rotten Tomatoes. For this exploratory data analysis, all movie titles are the same.

The point of contention is to find out whether Fandango's movie ratings are skewed more positive compared to the others. This arose because of the conflict of interests that Fandango also sells movie tickets for a commission. This was picked up by news outlet, 538, and the author used data analysis to support his findings. This exercise is to confirm his findings.

Here are the steps guided by Jose Portilla's tutorial:

  1. Read two open-sourced data from 538's repository.
  2. Explore the DataFrame properties.
  3. Explore the relationship between the popularity of a film and its rating.
  4. Create a scatterplot showing the relationship between rating and votes.
  5. Calculate the correlation between columns.
  6. Visualise the count of movies per year featured in Fandango with a plot. (This is to validate why the year 2015 was chosen since it was the largest data set available then.)
  7. Plot KDE to show the distribution and set a lower and upper cap to contain 0 to 5 ratings.
  8. Repeat steps for the other three providers.
  9. Merge DataFrames.
  10. Compare their KDE.
  11. Conclude that KDE for Fandango is skewed towards higher ratings compared to the others.
]]><![CDATA[Section 3: Seaborn]]>https://nurfadhilah.tech/section-3-seabornhttps://nurfadhilah.tech/section-3-seabornSun, 19 Sep 2021 15:40:18 GMTHere are the references I used:

5_ml_3_19Sep21.png

]]><![CDATA[Section 2: Matplotlib]]>https://nurfadhilah.tech/section-2-matplotlibhttps://nurfadhilah.tech/section-2-matplotlibSat, 18 Sep 2021 15:57:27 GMTHere are the resources I used:

  • https://matplotlib.org
  • https://www.udemy.com/course/python-for-machine-learning-data-science-masterclass

2_ml_1a_16Sep21.png

]]><![CDATA[Section 1: Pandas (part 2 of 2)]]>https://nurfadhilah.tech/section-1-pandas-part-2-of-2https://nurfadhilah.tech/section-1-pandas-part-2-of-2Fri, 17 Sep 2021 15:35:46 GMTClick here for part 1

Here's part two of the notes:

ml_1b_17Sep21.png

]]><![CDATA[Section 1: Pandas (part 1 of 2)]]>https://nurfadhilah.tech/section-1-pandas-part-1-of-2https://nurfadhilah.tech/section-1-pandas-part-1-of-2Thu, 16 Sep 2021 15:41:49 GMTThe documentation for Pandas is very comprehensive, you may wish to browse through it for more examples.

References used:

ml_1_16Sep21.png

]]><![CDATA[Section 0: NumPy]]>https://nurfadhilah.tech/section-0-numpyhttps://nurfadhilah.tech/section-0-numpyWed, 15 Sep 2021 13:49:39 GMTA one-page summary on NumPy. This is non-exhaustive but those listed are commonly used.

References:

ml_0_15sep21.png

]]>