Today's Progress: Spent most of the day reading up on regression. trying to prove autocorrelation, collinearity and creating graphs for the presentation.
Thoughts: Luckily i have Kat on my team. this short sprint would of suck if not for her.
Link(s) to work
Today's Progress: Completed all the models, focus on SVR especially the rbf and poly kernels. helped Apollo team member with a function that was really interesting.
Thoughts: need to get back to codewars, or writing summary articles.
Link(s) to work
Today's Progress: completed presentation slides, performed presentation and completed the report.
Thoughts: luckily i had Kat or this sprint would of been an epic fail.
Link(s) to work
Today's Progress: completed chapter 4 of DS in Python
Thoughts: I will complete this book, but first focus on on classifiers
Link(s) to work
Today's Progress: Chapter 2 of Intro to ML with Python.Supervised Machine Learning section. went over k-nearest neighbours and Linear Models.
Thoughts: I will start a notebook with a summary of each algorith, the benefits and when to use. as well as sample code and pipelines.
Link(s) to work
Today's Progress: Wrote summaries of k-NN and Decision Trees as well as Ensemble methods
Thoughts:
Link(s) to work
Today's Progress: completed Network analysis thereby completing the datsci track of DataCamp!!
Thoughts: It is a really good track. found it very useful, but revision is key as I forgotten alot of it. Luckily it is easier to review than learn from scratch :)
Today's Progress: Went through Linear models and SVMs in supervised learning in Intro to ML with Python
Thoughts:
Link(s) to work
Today's Progress: did two problems in CodeWars. Went over GridSearch, cross-validation and Pipeline.
Thoughts: I am hoping to master GridSearch and Pipeline to speed up modeling.
Link(s) to work
Today's Progress: summarizing Linear models for regression and classifying. Started the NLP in python course
Thoughts:
Link(s) to work
Today's Progress: finished chapter on sentiment analyzer
Thoughts:
Link(s) to work
Today's Progress: Started the NLP course in Udemy
Thoughts:
Link(s) to work
Today's Progress: Completed NLP in Python from DataCamp
Thoughts: It'sa very good intro course. There's alot to learn on the topic. I need to find a thorough course on NLP
Link(s) to work
Today's Progress: Completed the NLP course in Udemy
Thoughts: This course doesn't leave one with much confidence in coding nlp techniques. will go over the nlp sections in other courses
Link(s) to work
Today's Progress: Completed NLP sections in both the DatSci AZ and DatSci ML Bootcamp.
Thoughts: I will start implementing some nlp processing tomorrow.
Link(s) to work
Today's Progress: Completed kNN, SVM and kernel SVM sections in both the DatSci AZ
Thoughts: starting with regex section in Colts Python3 Udemy course.
Link(s) to work
Today's Progress: Completed the RegEx section of Colt. Completed the Linear Classifiers in Python course in DataCamp
Thoughts: Tomorrow I will put my knew regex knowlede to preprocess the mbti text dataset.
Link(s) to work
Today's Progress: I wrote all the individual functions for creating each feature in the dataset, then combined them all in a single function. regex really usefull.
Thoughts: I found a great blog that had regex exercises. will include them in my summary to drill the concepts into memmory
Link(s) to work
Today's Progress: Finished summary of kernel SVM, worked on plotly charts; bar charts.
Thoughts: west through the layout lectures of Dash. it is not complocated, although the syntax is intimidating with all the nested dictionaries.
Link(s) to work
Today's Progress: Completed the rest of the plotly charts and moving on to actual Dash. Completed the third section of Time Series analyses of DataCamp.
Thoughts:
Link(s) to work
Today's Progress: went over Dash layout, html and core components. Created my own stopwords list for the Personality profile sprint. and started thinking about pipelines and working out a strategy.
Thoughts: Dash is very interesting. will look into hosting the server on raspi, as well as finding a great project to practive what i have learned.
Link(s) to work
Today's Progress: Did the baseline result using countvectorizer and multinomialNB. played with a few parameters and results were dismal. Researched FeatureUnion, functionTransformer to join my created features along with countvectorizer to train the model.
Thoughts: I am probably going well above what's expected but learning about featureUnion is helpful in expanding my knowledge.
Link(s) to work
Today's Progress: I built functions to get numerical and text data used functionTransformer to be able to use them in a pipeline. built text and numerical pipeline and merged them with FeatureUnion into main pipeline
Thoughts: after successfully creating main pipeline, had errors with the text part after creating y_labels (therefore creating binary labels)
Link(s) to work
Today's Progress: struggled half the day trying to solve the main pipeline issue. gave up after mega headache.
Thoughts: starting over, concentrating on just nlp for the model. then include numerical features.
Link(s) to work
Today's Progress: not feeling very well, went over Dash and started a new course on algorithms and interview questions
Thoughts:
Link(s) to work
Today's Progress: went over Dash, built an interactive display using a range slider and outputting to a div.
Thoughts: I'm thinking of reviving the dc enviro project using dash instead of highcharts. I'm using flask anyway.
Link(s) to work
Today's Progress: got a working model using Random Forest. basic settings, need to do gridsearch on hyperparameters.
Thoughts:
Link(s) to work
Today's Progress: started over building a function to run through all the pipelines and taking into account the time to train. Next is working on gridsearch to fine tune the parameters.
Thoughts:
Link(s) to work
Today's Progress: working hard on the personality sprint project. all fell apart but was able to plan a b c d it and managed to complete and meet the deadline!
Thoughts: I learned so much this sprint but prolly performed the worst ironically...
Link(s) to work
Today's Progress: Spent most of the day prepping the presentation, creating wordcloud, and gaining insights into the data
Thoughts:
Link(s) to work
Today's Progress: Presentation day! went well. last minute prep and nlp insights gathering really helped. Started The Tensorflow Udemy course and finished the intro.
Thoughts: I was exhausted, individual project stress took its toll. then two of us carrying the rest of the team meant added pressure and stress. glad it's over.
Link(s) to work
Today's Progress: Nearly completed Tensorflow basicss and starting the exercises.
Thoughts: The estimator api seems really simple. will implement the previous classification problem using tensorflow estimator api.
Link(s) to work
Today's Progress: Completed Tensorflow basics and completed the reg and classification exercises.
Thoughts:
Today's Progress: Started on CNN, convoluted neural networks. need to go over the example.
Thoughts: I started exploring the movies dataset and need to plan the layout of the dashboard.
Today's Progress: completed the first two sections of Kaggles Learn Pandas section; reading and creating df's, and indexing and selecting.
Thoughts: I plan to complete all 7 by the end of review week
Today's Progress: I've gone through the CNN section of the tensorflow course. gone through the exercise solution. read ISLR chapter 8; trees and ensembles
Thoughts: I hope to build an initial model early on, so that i can trial others and iterate through to the winning model
Today's Progress: I completed 27 exercises on functions and another 3 on dictionaries
Thoughts: picked up cool tricks,
Today's Progress: I completed all exercises on dictionaries, wrote a notebook for reference and reviewed lists and sets
Thoughts: I will make short posts on lessons and other examples with which to apply them.
Today's Progress: Reviewed advance features of lists, sets and dictionaries.
Thoughts:
Today's Progress: Started another DL course 'Deep Learning A-Z' on Udemy, went through the first chapter.
Thoughts: Using Keras is SO much simpler that TF. I think learning and aplying Keras will be better to get started using DL, and then learn TF proper in case i need to go deeper.
Today's Progress: I completed the Forward feed NN chapter from DL A-Z, learned to crossval and gridsearch using the keraswrapper and scikit-learn.
Thoughts: I need to find a few basic datasets to test my usage of coding a NN. maybe from the datasets used in ML A-Z course.
Today's Progress: Gone over Decision Trees and Random Forest from ML A-Z.
Thoughts: Reading an article on embedding ML in a webapp i came across batching huge data and training the model sequntially using .partial_fit(). waiting on the next article where the writer uses flask to run the webapp.
Today's Progress: I spent most of the day reviewing kaggle kernels on predicting games.
Thoughts:
Today's Progress: Built a basic model to predict results. Then refactored it into a function.
Thoughts: Now that I have a basic model, I can build one with more vairables and hopefully more accurate.
Today's Progress: terrible migraine. not much work done
Thoughts:
Today's Progress: None.
Thoughts: Eid. took a well deserve break and spent the day with family and friends.
Today's Progress: Did most of the Baye Test Due tommorow night. Went over the intuition for word2vec.
Thoughts: Found Sendex tutorial on Dash, checking what he has to add
Today's Progress: Fathers day, gone for half the day. but i was left alone for few hours to complete the rest of the test.
Thoughts:
Today's Progress: Completed the Bayes test 70% -not great, but ok. Got my AWS working, but we using a colleagues as the central db to work from. Went over the train document on building a world cup model and using simulation.
Thoughts:
Today's Progress: I completed the pythonhow challenge. instructions abit unclear, but with some motivation from the discord chat i completed it.
Thoughts: finding it difficult to power through building a model. going to rest tonight and make sure i have a working one tomorrow night
Today's Progress:
Thoughts: review pending
Today's Progress: mastering sql
Thoughts:
Today's Progress: leaning api
Thoughts:
Today's Progress: learning selenium and apis
Thoughts:
[ Template ]
Today's Progress:
Thoughts:
Link(s) to work