Improve performance by making sure [ujson](https://pypi.python.org/pypi/ujson) is integrated everywhere in the stack it should be. - [ ] incoming [requests](https://github.com/requests/requests/issues/1595) - [ ] tinydb connections - [ ] pandas [to_json](https://github.com/pandas-dev/pandas/blob/master/pandas/io/json/json.py#L18) - `pandas.io.json.dumps = ujson.dumps` - `pandas.io.json.loads = usjon.loads` + `precise_float=True` in `pd.read_json()` args - [ ] skip `Flask.jsonify` by building response directly from pandas
Improve performance by making sure ujson is integrated everywhere in the stack it should be.
pandas.io.json.dumps = ujson.dumpspandas.io.json.loads = usjon.loads+precise_float=Trueinpd.read_json()argsFlask.jsonifyby building response directly from pandas