Normalize model quantities

[dp-rice]

Resolves #77. Normalizes input data so that log parameters are centered around zero. Makes it easier to specify prior distributions. Also move posterior plotting to fit.py instead of a Jupyter notebook.

$ ./fit.py 2> /dev/null

------------------------------------------------------------
Crits_Christoph: sars_cov_2 relative abundance per incidence
------------------------------------------------------------
Posterior arithmetic mean:
2.2e-11
Posterior geometric mean:
1.8e-11
Posterior quantiles:
     5%        25%        50%        75%        95%
7.7e-12    1.2e-11    1.7e-11    2.5e-11    4.7e-11

----------------------------------------------------
Rothman: sars_cov_2 relative abundance per incidence
----------------------------------------------------
Posterior arithmetic mean:
6.0e-12
Posterior geometric mean:
5.9e-12
Posterior quantiles:
     5%        25%        50%        75%        95%
4.1e-12    5.1e-12    5.9e-12    6.8e-12    8.5e-12

-----------------------------------------------------------
Crits_Christoph: norovirus relative abundance per incidence
-----------------------------------------------------------
Posterior arithmetic mean:
1.2e-10
Posterior geometric mean:
9.8e-11
Posterior quantiles:
     5%        25%        50%        75%        95%
3.6e-11    6.5e-11    9.8e-11    1.5e-10    2.6e-10

---------------------------------------------------
Rothman: norovirus relative abundance per incidence
---------------------------------------------------
Posterior arithmetic mean:
1.1e-08
Posterior geometric mean:
1.1e-08
Posterior quantiles:
     5%        25%        50%        75%        95%
8.3e-09    9.9e-09    1.1e-08    1.2e-08    1.5e-08