Monday, December 28, 2020

 

Big picture view is all about getting an overview of the system. Understanding what components make up the system, their interactions with each other and the external world. Easier said than done, if your strategy is to passively wait to get exposed to each part of the system and then build your overall system knowledge. However, there is a subtle framework called DSM, which by the way MIT dotes on, that can help us do it and get us ahead of the game faster.

 

The Design Structure Matrix methodology

 

To build the DSM framework for a given design, project plan or organization chart, we need 2 simple steps:

- breakdown the system under analysis into smaller elements

- draw out all the information flows between these elements

 

And then, equipped with above chart, we fill a table that would look like this



Diagonal boxes are mostly to be ignored. Look out for the boxes which are marked with a dot or a cross or whatever suits your aesthetic sense. The blank boxes essentially indicate no dependencies.

Rows – If we go across any row, the marked boxes indicate which column elements in the matrix have a dependency on the row element. Eg. Element B provides some information on which elements A, C, D, F, H and I are dependent on. Similarly, Element F provides some information on which elements B and C are dependent on.

Columns – If we go down a column, the marked boxes indicate which row elements in the matrix the column element is dependent on. Eg. Element E is dependent on information from elements C, D, G, H and I.

Closing Note

As we can see, the DSM provides a great visual view of the various interplay between all elements that make up your system. Giving this kind of view to the teams will help them to indulge in what we call Systems Thinking and enable them to think creatively. You can apply this framework on any kind of engineering or managerial problem. Use it and reap the benefits!


Reference: Applying the Design Structure Matrix to System Decomposition and Integration Problems: A Review and New Directions - Tyson R. Browning

Tuesday, December 22, 2020

Pandas - Series

 Went through Series data structure. It is similar to what one would want to do with lists or dictionaries but this data structure makes it more malleable to handle large data and makes computation much more faster.

Sunday, June 4, 2017

Nearest Neighbor

Analyzed:
http://scikit-learn.org/stable/modules/neighbors.html

>>> X = np.array([[-1,-1], [-2,-1], [-3,-2], [1,1], [2,1], [6,3]])
>>> nbrs = NearestNeighbors(n_neighbors=2, algorithm='ball_tree').fit(X)
>>> distances, indices = nbrs.kneighbors(X)
>>> indices
array([[0, 1],
       [1, 0],
       [2, 1],
       [3, 4],
       [4, 3],
       [5, 4]], dtype=int64)
>>> distances
array([[ 0.        ,  1.        ],
       [ 0.        ,  1.        ],
       [ 0.        ,  1.41421356],
       [ 0.        ,  1.        ],
       [ 0.        ,  1.        ],
       [ 0.        ,  4.47213595]])
>>> X = np.array([[-1,-1], [6,6], [-2,-1], [-3,-2], [1,1], [2,1], [6,3]])
>>> nbrs = NearestNeighbors(n_neighbors=2, algorithm='ball_tree').fit(X)
>>> distances, indices = nbrs.kneighbors(X)
>>> indices
array([[0, 2],
       [1, 6],
       [2, 0],
       [3, 2],
       [4, 5],
       [5, 4],
       [6, 1]], dtype=int64)
>>> distances
array([[ 0.        ,  1.        ],
       [ 0.        ,  3.        ],
       [ 0.        ,  1.        ],
       [ 0.        ,  1.41421356],
       [ 0.        ,  1.        ],
       [ 0.        ,  1.        ],
       [ 0.        ,  3.        ]])
>>> nbrs.kneighbors_graph(X).toarray()
array([[ 1.,  0.,  1.,  0.,  0.,  0.,  0.],
       [ 0.,  1.,  0.,  0.,  0.,  0.,  1.],
       [ 1.,  0.,  1.,  0.,  0.,  0.,  0.],
       [ 0.,  0.,  1.,  1.,  0.,  0.,  0.],
       [ 0.,  0.,  0.,  0.,  1.,  1.,  0.],
       [ 0.,  0.,  0.,  0.,  1.,  1.,  0.],
       [ 0.,  1.,  0.,  0.,  0.,  0.,  1.]])

Sunday, April 30, 2017

Reading prostate dataset into a variable

I have installed anaconda and created an environment where I have installed all statistics related libraries like NumPy, SciPy, Pandas.


On the ESL website: https://statweb.stanford.edu/~tibs/ElemStatLearn/, I goto Data tab and open the prostate dataset. I then copy this dataset onto an excel (eg. 'prostate.xls'). Note: Each column should be aligned and separated.


I then copy this excel into the C:\ProgramData\Miniconda2\envs\tutorial folder, wherein lies the python.exe.


I then start the application and run the below commands
>>> import pandas
>>> df = pandas.read_excel('prostate.xls')



Essential Links


Dec-2020
https://www.coursera.org/learn/python-data-analysis/home/welcome